Temperature energy-saving control method and system for communication machine room

Abstract

The invention relates to the technical field of communication machine room energy conservation, in particular to a temperature energy-saving control method and system for a communication machine room, and the method comprises the steps: obtaining multi-region temperature and humidity and equipment load data in the machine room, inputting a pre-training model to obtain a cabinet heat dissipation risk level, judging an environment state through combination with an incidence matrix, entering an operation mode if a requirement is satisfied, and entering a control mode if a requirement is satisfied. If not, a local temperature control adjusting mechanism is triggered; the method comprises the following steps: calculating a refrigeration efficiency ratio according to the air volume of ventilation equipment, air conditioner power consumption, real-time electricity price and the like, generating a candidate scheme set by using a dynamic weight multi-target optimization algorithm in combination with factors such as equipment life, and selecting a target scheme through virtual verification; and according to air conditioner waste heat and cabinet heat dissipation and storage heat, a heat energy distribution strategy is dynamically generated by using a deep Q network reinforcement learning algorithm. Therefore, the problems that in the prior art, energy is wasted, cost is high, heat dissipation requirements of all areas are difficult to accurately adapt, the service life and stability of equipment are affected, and fault risks are increased are solved.

Description

Technical Field

[0001] This application relates to the technical field of energy conservation in communication computer rooms, and particularly to a temperature energy conservation control method and system for a communication computer room. Background Art

[0002] With the rapid development of information technology, the scale and quantity of communication computer rooms are constantly increasing. A large amount of heat is generated by the continuous operation of communication equipment in the computer room. To ensure the normal operation of the equipment, it is necessary to maintain a suitable temperature and humidity environment through refrigeration equipment.

[0003] However, traditional temperature control methods for communication computer rooms often adopt fixed refrigeration strategies, such as always keeping the air conditioner running at full power. This method does not fully consider factors such as the actual load changes in the computer room, outdoor environmental conditions, and the heat dissipation characteristics of equipment, resulting in serious energy waste and high operating costs. At the same time, the equipment loads in different areas of the computer room are different, and the heat distribution is uneven. The traditional unified temperature control strategy is difficult to accurately meet the heat dissipation requirements of equipment in each area, which not only affects the service life and operation stability of the equipment, but also increases the risk of equipment failure. Summary of the Invention

[0004] This application provides a temperature energy conservation control method and system for a communication computer room to solve problems such as energy waste, high cost, difficulty in accurately adapting to the heat dissipation requirements of each area, affecting the equipment life and stability, and increasing the risk of failure in the prior art.

[0005] The first aspect of the embodiments of this application provides a temperature energy conservation control method for a communication computer room, including the following steps: obtaining temperature data, humidity data, and communication equipment load data of multiple areas in the communication computer room; inputting the temperature data, humidity data, and communication equipment load data into a pre-trained equipment health assessment model, outputting the heat dissipation risk level of the cabinet, and combining a preset communication equipment operating temperature-humidity-load correlation matrix to determine whether the environmental status of each area in the computer room meets the operating requirements. If it meets the operating requirements, enter the operating mode; otherwise, trigger a local temperature control adjustment mechanism; after entering the operating mode or triggering the local temperature control adjustment mechanism, calculate the refrigeration efficiency ratio based on the air volume of the ventilation equipment, air conditioner power consumption, and real-time electricity price data, and combine the remaining service life and performance attenuation coefficient of the refrigeration equipment to generate a candidate control solution set through a dynamic weight multi-objective optimization algorithm. Perform virtual verification on the candidate control solution set, and select the solution with the highest comprehensive score as the target solution; based on the target solution, store heat according to the waste heat of the air conditioner and the heat dissipation of the cabinet, and use the deep Q-network reinforcement learning algorithm to dynamically generate a heat energy distribution strategy.

[0006] Optionally, if the running requirements are met, enter the running mode, including: obtaining outdoor meteorological data, where the outdoor meteorological data includes temperature, humidity, wind speed, etc.; if the outdoor temperature is less than or equal to the first preset temperature, turn off the air conditioner and start natural ventilation. At the same time, according to the wind speed and wind direction data, dynamically optimize the opening degree of the ventilation opening and the angle of the deflector through a fluid dynamics simulation model; if the outdoor temperature is greater than the first preset temperature and less than or equal to the second preset temperature, turn off the air conditioner and start natural ventilation, and dynamically adjust the operating intensity of the air conditioner and ventilation equipment based on the temperature and humidity difference between indoors and outdoors, the real-time temperature and humidity data in the computer room, and the load condition of communication equipment through a fuzzy control algorithm; if the outdoor temperature is greater than the second preset temperature, fully operate the air conditioner in advance, and calculate the optimal combination of the air outlet angles of the air conditioner group according to the hot spot distribution in the computer room through an ant colony optimization algorithm.

[0007] Optionally, trigger a local temperature control adjustment mechanism, including: adjusting the blade angle of the air supply floor according to the cabinet temperature data, and real-time monitoring the change trend of the cabinet temperature data. When it is found that the cabinet temperature continues to rise and reaches the preset condition, start the local auxiliary refrigeration equipment.

[0008] Optionally, the preset condition is that the difference between the cabinet temperature and the average temperature of the computer room exceeds 5°C, or the cabinet temperature reaches the preset temperature threshold.

[0009] Optionally, use the deep Q-network reinforcement learning algorithm to dynamically generate a thermal energy distribution strategy, including: obtaining the remaining capacity of the thermal energy storage device, the predicted value of the thermal energy demand, and the priority of communication equipment; integrating the remaining capacity of the thermal energy storage device, the predicted value of the thermal energy demand, and the priority of communication equipment, and inputting the temperature and humidity and equipment load information of each area of the computer room into the deep Q-network reinforcement learning algorithm. Determine the target action through the Q value output by the network and the ε-greedy strategy; execute the target action, obtain a new state and reward from the environment, store the current state, action, reward, and new state as an experience tuple in the experience replay buffer, and randomly sample experience tuples from the experience replay buffer. Update the weights of the deep Q-network according to the reward and the target Q value, dynamically adjust the policy parameters, evaluate and optimize the policy, and generate the target thermal energy distribution strategy.

[0010] Optionally, after generating the target thermal energy distribution strategy, it includes: obtaining the operating parameters of the thermal energy distribution; establishing a fault diagnosis model, inputting the operating parameters into the fault diagnosis model to determine whether there is a fault. If there is no fault, proceed with the operation. Otherwise, automatically analyze the cause and location of the fault, and generate detailed repair suggestions; automatically adjust the thermal energy distribution strategy according to the modification suggestions.

[0011] Optionally, perform virtual verification on the set of candidate control schemes, and select the scheme with the highest comprehensive score as the target scheme, including: constructing a virtual model for temperature energy-saving control of a communication machine room based on digital twin technology; inputting the target scheme into the virtual model for operation to obtain the energy-saving effect of the scheme, the operation stability of the equipment, and the impact on the operating environment of communication equipment. If the simulation results do not meet the expected standards, iterate and optimize the target scheme.

[0012] Optionally, after obtaining the temperature data, humidity data, and communication equipment load data of multiple areas in the communication machine room, it further includes: obtaining historical data; using time series analysis algorithms to model the historical data to predict the temperature and humidity change trends of each area in the machine room and the change of the communication equipment load; according to the prediction results, adjust the operation mode or the parameters of the local temperature control adjustment mechanism in advance to perform predictive maintenance on the refrigeration equipment and ventilation equipment.

[0013] Optionally, the formula for the equipment health assessment model is: where is the health assessment value of the i-th cabinet, is the current temperature of the i-th cabinet, is the current humidity of the i-th cabinet, is the current load value of the i-th cabinet, is the maximum allowable threshold of temperature, is the maximum allowable threshold of humidity, is the maximum allowable threshold value of the load value, is the temperature weight coefficient, is the humidity weight coefficient, is the load value weight coefficient.

[0014] The second aspect of the present application provides a temperature energy-saving control system for a communication computer room, including: an acquisition module for acquiring temperature data, humidity data, and communication device load data of multiple areas in the communication computer room; a judgment module for inputting the temperature data, humidity data, and communication device load data into a pre-trained device health assessment model, outputting the heat dissipation risk level of the cabinet, and combining a preset communication device operation temperature-humidity-load association matrix to judge whether the environmental status of each area in the computer room meets the operation requirements. If it meets the operation requirements, it enters the operation mode; otherwise, it triggers a local temperature control adjustment mechanism; a calculation module for, after entering the operation mode or triggering the local temperature control adjustment mechanism, calculating the refrigeration efficiency ratio based on the air volume of the ventilation equipment, air-conditioning power consumption, and real-time electricity price data, and combining the remaining service life and performance decay coefficient of the refrigeration equipment, generating a candidate control solution set through a dynamic weight multi-objective optimization algorithm, virtually verifying the candidate control solution set, and selecting the solution with the highest comprehensive score as the target solution; a generation module for, based on the target solution, storing heat according to the waste heat of the air conditioner and the heat dissipation of the cabinet, and dynamically generating a heat energy distribution strategy by using a deep Q-network reinforcement learning algorithm.

[0015] Therefore, the present application has at least the following beneficial effects: In the embodiments of the present application, the temperature, humidity, and communication device load data of each area in the computer room are collected in real time through a multi-source sensor network. The heat dissipation risk level is quantified by a pre-trained device health assessment model, and the environmental adaptability dynamic diagnosis is realized by combining the device operation temperature-humidity-load association matrix, accurately identifying the areas with excessive environment and triggering the local temperature control response mechanism to avoid the "one-size-fits-all" energy consumption waste of traditional unified regulation. Secondly, by integrating the refrigeration efficiency ratio calculation model and the dynamic weight multi-objective optimization algorithm, with the real-time electricity price, the remaining life of the equipment, and the performance decay coefficient as constraint variables, a candidate solution set that takes into account cost, efficiency, and reliability is generated through virtual simulation verification, and the optimal strategy is selected based on multi-dimensional comprehensive scoring, breaking the decision-making dilemma of the separation of cost and performance in the generation of traditional solutions. The mechanism of coordinated utilization of air conditioner waste heat recovery and cabinet heat dissipation is introduced, and a dynamic heat energy distribution model is constructed by combining the deep Q-network reinforcement learning algorithm. Through iterative optimization of reinforcement learning, intelligent decision-making for heat energy storage, transfer, and reuse is realized, and low-efficiency waste heat is converted into high-efficiency energy supply resources. Therefore, the problems of energy waste, high cost, difficulty in accurately adapting to the heat dissipation requirements of each area, affecting the life and stability of equipment, and increasing the risk of failure in the prior art are solved.

[0016] The additional aspects and advantages of the present application will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present application. Description of the Drawings

[0017] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of embodiments in conjunction with the accompanying drawings, where: Figure 1 FIG. is a flowchart of a temperature energy-saving control method for a communication machine room according to an embodiment of the present application; Figure 2 FIG. is a schematic diagram of the temperature change in the core server area A according to an embodiment of the present application; Figure 3 FIG. is a schematic diagram of the refrigeration efficiency ratios of different schemes according to an embodiment of the present application; Figure 4 FIG. is a block diagram example of a temperature energy-saving control system for a communication machine room according to an embodiment of the present application; Figure 5 FIG. is a schematic structural diagram of an electronic device according to an embodiment of the present application; Among them, 100 - acquisition module, 200 - judgment module, 300 - calculation module, 400 - generation module. Detailed Embodiments

[0018] Embodiments of the present application will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals indicate the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present application and should not be construed as limiting the present application.

[0019] The temperature energy-saving control method and system for a communication machine room according to an embodiment of the present application will be described below with reference to the accompanying drawings. In view of the problem mentioned in the above background art that it is difficult to accurately adapt to the heat dissipation requirements of each area, the present application provides a temperature energy-saving control method for a communication machine room. In this method, the temperature, humidity, and communication device load data of each area in the machine room are collected in real time through a multi-source sensor network, the heat dissipation risk level is quantified by a pre-trained device health assessment model, and the environmental adaptability dynamic diagnosis is realized by combining the device operation temperature-humidity-load correlation matrix, accurately identifying the areas where the environment exceeds the standard and triggering a local temperature control response mechanism to avoid the "one-size-fits-all" energy consumption waste of traditional unified regulation. Secondly, by integrating the refrigeration efficiency ratio calculation model and the dynamic weight multi-objective optimization algorithm, with the real-time electricity price, the remaining life of the device, and the performance decay coefficient as constraint variables, a candidate solution set that takes into account cost, efficiency, and reliability is generated through virtual simulation verification, and the optimal strategy is selected based on multi-dimensional comprehensive scoring, breaking the decision-making dilemma of the separation of cost and performance in the generation of traditional solutions. The mechanism of coordinated utilization of air-conditioning waste heat recovery and cabinet heat dissipation is introduced, and a dynamic heat energy distribution model is constructed by combining the deep Q-network reinforcement learning algorithm. Through reinforcement learning iterative optimization, the intelligent decision-making of heat energy storage, transfer, and reuse is realized, and the low-efficiency waste heat is converted into high-efficiency energy supply resources. Thus, the problems in the prior art such as energy waste, high cost, difficulty in accurately adapting to the heat dissipation requirements of each area, affecting the service life and stability of devices, and increasing the risk of failures are solved.

[0020] The temperature energy-saving control method and system for a communication machine room according to an embodiment of the present application will be described below with reference to the accompanying drawings.

[0021] Specifically, Figure 1 is a schematic flowchart of a temperature energy-saving control method for a communication machine room provided by an embodiment of the present application.

[0022] As Figure 1 shown, the temperature energy-saving control method for the communication machine room includes the following steps: In step S101, temperature data, humidity data, and communication device load data of multiple areas in the communication machine room are obtained.

[0023] Among them, the communication machine room refers to an enclosed space where communication devices (such as servers, switches, base station controllers, etc.) are centrally deployed. Usually, the machine room contains multiple functional partitions, and there are significant differences in device loads in different areas; the communication device load data refers to the resource occupancy situation under the real-time operation state of the device, including but not limited to power load, computing load, network load, and storage load.

[0024] It can be understood that, by obtaining the temperature data, humidity data, and communication device load data of multiple areas in the communication machine room in the embodiment of the present application, a comprehensive and accurate data basis is provided for subsequent analysis.

[0025] In an embodiment of the present application, after obtaining the temperature data, humidity data and communication equipment load data of multiple areas in the communication room, it also includes: obtaining historical data; using a time series analysis algorithm to model the historical data, and predicting the temperature and humidity change trends in each area of the room and the changes in the communication equipment load; according to the prediction results, adjusting the operating mode or the parameters of the local temperature control mechanism in advance, and performing predictive maintenance on the refrigeration equipment and ventilation equipment.

[0026] Among them, historical data refers to various operating data continuously recorded and stored in the communication room over the past period of time, including but not limited to the temperature, humidity, communication equipment load, operating parameters of air conditioning and ventilation equipment, energy consumption data, etc. of each area.

[0027] It is understandable that the embodiments of the present application can prospectively grasp the temperature, humidity fluctuations and load trends by mining historical data and using time series analysis algorithms to model and predict changes in the computer room environment and equipment load. Based on accurate predictions, the operating mode and temperature control mechanism parameters can be dynamically adjusted in advance, such as optimizing the air conditioning operation strategy before high temperatures arrive, pre-adjusting the power of ventilation equipment according to load changes, and realizing refined allocation and efficient utilization of energy, effectively reducing energy consumption costs. At the same time, predictive maintenance of refrigeration and ventilation equipment can accurately locate potential failure risks, and timely intervene in processing before equipment performance declines or failures occur, avoiding business interruptions caused by sudden downtime.

[0028] In step S102, the temperature data, humidity data, and communication equipment load data are input into the pre-trained equipment health assessment model, the heat dissipation risk level of the cabinet is output, and combined with the preset communication equipment operating temperature-humidity-load association matrix, it is determined whether the environmental status of each area in the computer room meets the operating requirements. If the operating requirements are met, the operating mode is entered; otherwise, the local temperature control adjustment mechanism is triggered.

[0029] Among them, the heat dissipation risk levels include low risk, medium risk and high risk. Among them, low risk means that the health value is ≤ the threshold of 60%, medium risk means that the threshold is 60%-80%; high risk means that the health value is ≥ the threshold of 80%.

[0030] It can be understood that in the embodiments of the present application, by inputting temperature, humidity, and equipment load data into a pre-trained model, the heat dissipation risk level of the cabinet is output, and the environmental state is judged in combination with the correlation matrix, so as to determine whether to enter the operating mode or trigger the local temperature control adjustment mechanism. This can accurately identify the environmental state of each area in the computer room, flexibly adjust the temperature control strategy according to the actual situation, and avoid the blindness of the traditional fixed cooling strategy. When the operating requirements are met, entering the operating mode can ensure the stable operation of the equipment in a suitable environment, while when the requirements are not met, triggering the local temperature control adjustment mechanism can timely solve problems such as local overheating, improve the energy utilization efficiency, reduce the operating cost, effectively increase the service life and operating stability of the equipment, and reduce the probability of failures.

[0031] It should be noted that the temperature-humidity-load correlation matrix is a matrix used to describe the mutual relationship among temperature, humidity, and the load of communication equipment in a communication computer room. The following is a temperature-humidity-load correlation matrix: Diagonal elements: The diagonal elements of the matrix are all 1, that is, the main diagonal elements a11 = a22 = a33 = 1. This indicates that the correlation between each variable and itself is a completely positive correlation, because the change trend of any variable itself must be completely consistent with itself, and the correlation coefficient of 1 is a mathematical definition and representation method.

[0032] Off-diagonal elements, temperature and humidity: a12 = a21 = 0.3, indicating that there is a positive correlation between temperature and humidity. When the temperature changes, the humidity has a certain co-directional change trend, but the degree of correlation is not very strong. This may be because when some equipment in the computer room operates to generate heat, it may also affect the water vapor content in the air, or when the temperature control system in the computer room adjusts the temperature, it will also have a certain indirect impact on the humidity.

[0033] Temperature and load: a13 = a31 = -0.5, indicating that there is a negative correlation between temperature and load. That is, when the load of communication equipment increases, the temperature has a downward trend. This may be because when the equipment load increases, the equipment itself dissipates more heat, so that the refrigeration equipment in the computer room needs to work harder to maintain the temperature, resulting in a relatively lower environmental temperature. Or it may be because the computer room adopts some intelligent temperature control strategies. When it detects an increase in equipment load, it will adjust the operating parameters of the refrigeration equipment in advance to prevent the temperature from being too high.

[0034] Humidity and Load: a23 = a32 = 0.2, indicating a weak positive correlation between humidity and load. Changes in load may cause some physical processes in the computer room to change, thereby having a certain impact on humidity, but this impact is relatively small. For example, when the equipment load increases, some chemical reactions or physical processes inside the equipment may release or absorb a certain amount of water vapor, resulting in a slight change in the humidity in the computer room.

[0035] In the embodiment of the present application, the formula of the equipment health assessment model is: Where, is the health assessment value of the i-th cabinet, is the current temperature of the i-th cabinet, is the current humidity of the i-th cabinet, is the current load value of the i-th cabinet, is the maximum allowable threshold of temperature, is the maximum allowable threshold of humidity, is the maximum allowable threshold of the load value, is the temperature weight coefficient, is the humidity weight coefficient, is the load value weight coefficient.

[0036] Specifically, the cabinets in the communication computer room need to be evaluated. Among them, the current temperature is 28 °C, the current humidity is 60%, the current load value is 75%. According to the industry standard of communication equipment, the maximum allowable value of temperature is 35 °C, the maximum allowable value of humidity is 80%, and the maximum allowable value of the load is 90%. If this cabinet is a high-temperature sensitive device, the temperature weight coefficient is 0.5, the humidity weight coefficient is 0.3, and the load value weight coefficient is 0.2. According to the formula of the equipment health assessment model: It can be obtained that is approximately 0.80. According to the preset risk threshold, this cabinet is in a high-risk state, and it is necessary to pay attention to the changes in temperature and load and adjust the ventilation or refrigeration strategy in advance.

[0037] In the embodiment of the present application, if the operation requirements are met, the operation mode is entered, including: obtaining outdoor meteorological data, where the outdoor meteorological data includes temperature, humidity, wind speed, etc.; if the outdoor temperature is less than or equal to the first preset temperature, the air conditioner is turned off and natural ventilation is started. At the same time, according to the wind speed and wind direction data, the opening degree of the ventilation opening and the angle of the deflector are dynamically optimized through a fluid dynamics simulation model; if the outdoor temperature is greater than the first preset temperature and less than or equal to the second preset temperature, the air conditioner is turned off and natural ventilation is started, and based on the temperature and humidity difference between indoor and outdoor, the real-time temperature and humidity data in the computer room, and the load condition of communication equipment, the operation intensity of the air conditioner and ventilation equipment is dynamically adjusted through a fuzzy control algorithm; if the outdoor temperature is greater than the second preset temperature, the air conditioner is pre-operated at full power in advance, and according to the hot spot distribution in the computer room, the optimal combination of the air outlet angles of the air conditioner group is calculated through an ant colony optimization algorithm.

[0038] Among them, the first preset temperature can be the threshold temperature for triggering natural ventilation to replace the air conditioner, such as 15°C, etc., and the second preset temperature can be the transition temperature threshold for combining the air conditioner and natural ventilation, such as 30°C, etc.

[0039] It can be understood that in the embodiment of the present application, by obtaining outdoor meteorological data such as temperature, humidity, and wind speed, a hierarchical response mechanism is constructed with the first and second preset temperatures as thresholds, and algorithms such as fluid dynamics simulation, fuzzy control, and ant colony optimization are deeply integrated to achieve intelligent coordination of natural ventilation and mechanical refrigeration. When the outdoor temperature is relatively low, the air conditioner is turned off and natural ventilation is enabled, and the ventilation structure is optimized in combination with fluid dynamics simulation to efficiently convert natural energy into heat dissipation power; when the temperature is moderate, the fuzzy control algorithm is used to comprehensively consider multiple parameters to dynamically adjust the operation intensity of the equipment, and an accurate balance is achieved between energy conservation and ensuring the equipment operation environment; when the temperature is high, the ant colony optimization algorithm is used to optimize the air outlet angle of the air conditioner according to the hot spot distribution, avoiding the energy waste and uneven refrigeration problems caused by traditional full-power operation. This not only greatly reduces the energy consumption of the computer room, but also from the perspective of the entire life cycle of equipment operation, reduces the loss of equipment caused by temperature fluctuations through accurate temperature control, improves the reliability and service life of the equipment, and effectively reduces the operation and maintenance costs at the same time.

[0040] Specifically, a certain communication computer room is located in a subtropical region, and the outdoor temperature changes greatly during the day in summer. At 8 am, the outdoor temperature is 12°C, the humidity is 60%, and the wind speed is 2 m / s. After obtaining these outdoor meteorological data, since the outdoor temperature is less than the first preset temperature of 15°C, the computer room turns off the air conditioner and starts natural ventilation. At the same time, according to the wind speed and wind direction data, the fluid dynamics simulation model calculates that when the opening degree of the ventilation opening is set to 50% and the angle of the deflector is adjusted to 30°, the air circulation efficiency in the computer room can be the highest. At this time, the computer room can maintain the appropriate operating temperature of the equipment through natural ventilation, avoiding the energy waste caused by the opening of the air conditioner.

[0041] At 12:00 noon, the outdoor temperature rises to 22°C, the humidity is 55%, and the wind speed is 1.5 m / s. At this time, the outdoor temperature is between the first preset temperature and the second preset temperature (30°C). The air-conditioning compressor in the computer room is turned off, the fan is kept running, and natural ventilation is started. Based on the indoor-outdoor temperature and humidity difference, the real-time temperature and humidity data in the computer room, and the load condition of communication equipment, the fuzzy control algorithm determines that the operating intensity of the ventilation equipment needs to be increased to 60%, and the air-conditioning fan speed is adjusted to 70%. While ensuring the normal operation of the equipment, energy consumption is effectively reduced.

[0042] At 3:00 pm, the outdoor temperature reaches 32°C, the humidity is 40%, and the wind speed is 0.8 m / s. The outdoor temperature is higher than the second preset temperature. The air-conditioning in the computer room runs at full power in advance. At the same time, through an infrared thermal imager, a hot spot appears in the cabinet area in the southwest corner of the computer room, and the temperature reaches 33°C. The ant colony optimization algorithm calculates the optimal combination of the air outlet angles of the air-conditioning group, and adjusts the air outlet angles of the corresponding air-conditioners to 15° and 25° respectively, so that the temperature in the hot spot area drops to 26°C within 15 minutes, ensuring the stable operation of communication equipment.

[0043] It should be noted that when dynamically optimizing the opening degree of the ventilation opening and the angle of the deflector through the fluid mechanics simulation model, the component form of the Navier-Stokes equation in the Cartesian coordinate system is as follows: Continuity equation: ; x-direction momentum equation: ; x-direction momentum equation: ; x-direction momentum equation: ; Among them, u, v, and w are the velocity components of the fluid in the x, y, and z directions respectively, ρ is the fluid density, p is the pressure, μ is the dynamic viscosity, and t is the time.

[0044] By solving the equations and combining the boundary conditions (such as the wind speed and wind direction of the ventilation opening), the flow field distribution under different opening degrees of the ventilation opening and angles of the deflector can be obtained, so as to optimize the ventilation structure.

[0045] By calculating the optimal combination of the air outlet angles of the air-conditioning group through the ant colony optimization algorithm, the probability that the ant selects the path (combination of air outlet angles) is: Ant The probability formula for selecting from node i to node j is: Among them, is the importance factor of pheromone, is the importance factor of heuristic information, is the heuristic information from node i to node j, is the ant A set of selectable nodes is the pheromone concentration on the path from node i to node j.

[0046] In the embodiment of the present application, triggering the local temperature control adjustment mechanism includes: adjusting the blade angle of the air supply floor according to the cabinet temperature data, and real-time monitoring the change trend of the cabinet temperature data. When it is found that the cabinet temperature continues to rise and reaches the preset condition, start the local auxiliary refrigeration equipment.

[0047] Among them, the preset condition is that the difference between the cabinet temperature and the average temperature of the computer room exceeds 5°C, or the cabinet temperature reaches the preset temperature threshold.

[0048] It can be understood that in the embodiment of the present application, by collecting the cabinet temperature data in real time and dynamically adjusting the blade angle of the air supply floor, the airflow distribution around the cabinet can be finely controlled, the local heat dissipation conditions can be targeted improved, and local high temperatures caused by airflow disorders can be avoided. At the same time, continuously monitoring the temperature change trend, when the cabinet temperature touches "the difference from the average temperature of the computer room exceeds 5°C" or "reaches the preset temperature threshold", quickly start the local auxiliary refrigeration equipment, realizing the accurate positioning and rapid response to the hot spots in the computer room, not only effectively reducing the energy consumption of traditional global refrigeration, but also significantly improving the operation stability of the equipment, extending the service life of the equipment, and reducing the risk of failures caused by overheating.

[0049] Specifically, in a communication computer room, the initial temperature of cabinet A is 26°C, which is lower than the average temperature of the computer room, 28°C. As the equipment runs, its temperature gradually rises to 29°C, and the system accordingly adjusts the corresponding blade angle of the air supply floor from 40° to 60° to strengthen heat dissipation. However, half an hour later, the temperature of cabinet A continues to climb to 36°C, exceeding the preset threshold of 35°C. At this time, the system immediately starts the small refrigeration fan beside the cabinet and cools the temperature to 32°C within 10 minutes to prevent the equipment from being damaged due to overheating.

[0050] In step S103, after entering the operation mode or triggering the local temperature control adjustment mechanism, based on the ventilation equipment air volume, air conditioner power consumption and real-time electricity price data, calculate the refrigeration efficiency ratio, and combine the remaining service life and performance decay coefficient of the refrigeration equipment, and generate a candidate control solution set through the dynamic weight multi-objective optimization algorithm. Perform virtual verification on the candidate control solution set, and select the solution with the highest comprehensive score as the target solution.

[0051] Among them, the candidate control solution set can be a set of possible control solutions generated by the dynamic weight multi-objective optimization algorithm.

[0052] It can be understood that in the embodiments of the present application, by comprehensively considering multi-dimensional factors such as the air volume of the ventilation equipment, the power consumption of the air conditioner, the real-time electricity price, the equipment life, and the performance attenuation, a refrigeration efficiency ratio evaluation system is constructed. Combining with the dynamic weight multi-objective optimization algorithm, a set of candidate control schemes is generated, breaking through the limitations of traditional single temperature control strategies. Using virtual verification technology to simulate the operation effects of each scheme, accurately evaluate the energy-saving benefits, equipment stability, and environmental adaptability, and finally screen out the target scheme with the optimal comprehensive score.

[0053] Specifically, in a certain communication computer room, the outdoor temperature is relatively high during the day in summer, and the air conditioner in the computer room starts to operate at full power. At this time, the sensor installed in the computer room collects the air volume of the ventilation equipment in real time, which is 8,000 cubic meters per hour, the power consumption of the air conditioner is 50 kilowatts, and the local real-time electricity price is 0.8 yuan per degree. Calculate the current refrigeration efficiency ratio based on these data. Assuming that the current refrigeration efficiency ratio is calculated to be 1.2 (unit cooling capacity / unit power consumption) through a specific formula.

[0054] At the same time, a certain main refrigeration equipment in the computer room has been used for 5 years. Through professional evaluation, its remaining service life is expected to be 3 years, and the performance attenuation coefficient is 0.7 (that is, the current performance is 70% of the initial performance).

[0055] Based on the above data, use the dynamic weight multi-objective optimization algorithm to generate a set of candidate control schemes: Scheme 1: Reduce the operating power of the air conditioner to 45 kilowatts, and at the same time increase the air volume of the ventilation equipment to 10,000 cubic meters per hour. It is expected that the refrigeration efficiency ratio can be increased to 1.3, which can reduce the power consumption to a certain extent, but may slightly increase the local temperature in the computer room, and has little impact on the remaining service life of the equipment. It is expected to extend the use by 1 month.

[0056] Scheme 2: Maintain the current power of the air conditioner, optimize the air outlet angle of the air conditioner, and replace some ventilation equipment filters to improve the ventilation efficiency. It is expected that the refrigeration efficiency ratio can be increased to 1.25, the remaining service life of the equipment is expected to be extended by half a month, and at the same time, the equipment performance attenuation speed can be reduced.

[0057] Scheme 3: Increase the power of the air conditioner to 55 kilowatts to strengthen the refrigeration effect and ensure the stability of the computer room temperature, but it will increase the power consumption, and the refrigeration efficiency ratio will drop to 1.1. However, it can reduce the performance attenuation of the equipment caused by high temperature, and it is expected that the remaining service life of the equipment will be extended by 2 months.

[0058] Then, a virtual model for temperature energy-saving control of a communication machine room is constructed based on digital twin technology, and the above three candidate solutions are respectively input into the virtual model for operation. During the simulation process, after 2 hours of operation of Solution 1, the temperature in some areas of the machine room exceeds the upper limit of the comfortable operating temperature of the equipment. Although the energy-saving effect is obvious, it has a certain impact on the operation stability of the equipment; during the simulation operation of Solution 2, the temperature and humidity in the machine room are stable, the equipment operates well, and the energy-saving effect and the improvement of equipment life are relatively balanced; although Solution 3 ensures the stability of the machine room temperature, the energy consumption increases significantly and the operating cost rises significantly.

[0059] Finally, according to indicators such as energy-saving effect, equipment operation stability, and impact on the operating environment of communication equipment, a comprehensive score is carried out. Solution 2 has the highest score and is selected as the target solution for application in the actual operation of the machine room, achieving the maximum optimization of energy consumption and equipment management on the premise of ensuring the stable operation of the machine room.

[0060] It should be noted that the formula for calculating the refrigeration efficiency ratio is: refrigeration efficiency ratio = refrigeration capacity / input power.

[0061] In the embodiment of the present application, virtual verification is carried out on the candidate control solution set, and the solution with the highest comprehensive score is selected as the target solution, including: constructing a virtual model for temperature energy-saving control of a communication machine room based on digital twin technology; inputting the target solution into the virtual model for operation to obtain the energy-saving effect, equipment operation stability, and impact on the operating environment of communication equipment of the solution. Among them, if the simulation result does not meet the expected standard, iterative optimization is carried out on the target solution.

[0062] Among them, digital twin technology is a technology that highly simulates physical entities using digital models. Through means such as data collection and analysis, modeling and simulation, a virtual model corresponding to the physical entity is constructed in the virtual space; the expected standards include that the energy-saving effect should reach a certain energy-saving rate, the equipment operation stability should ensure that the fault-free operation time of the equipment reaches a certain standard, and the impact on the operating environment of communication equipment should ensure that various environmental parameters are kept within the range allowed by the equipment, etc.

[0063] It can be understood that the embodiment of the present application constructs a virtual model for temperature energy-saving control of a communication machine room with the help of digital twin technology, runs and simulates the target solution in a virtual environment, can intuitively present the effects of the solution in terms of energy saving, stable operation of equipment, and environmental impact, and through comparison with the expected standard, iterative optimization is carried out on the unqualified solutions, realizing efficient verification and improvement of the energy-saving solution before actual application and effectively reducing the trial-and-error cost.

[0064] In step S104, based on the target solution, heat storage is carried out according to the waste heat of the air conditioner and the heat dissipation of the cabinet, and a thermal energy distribution strategy is dynamically generated using the deep Q-network reinforcement learning algorithm.

[0065] It can be understood that the embodiments of the present application are based on the target solution, integrating the waste heat of the air conditioner and the heat dissipation of the cabinet for heat storage, and using the deep Q-network reinforcement learning algorithm to dynamically generate the heat energy distribution strategy, efficiently utilize the heat energy resources, and improve the energy utilization efficiency.

[0066] In the embodiments of the present application, the deep Q-network reinforcement learning algorithm is used to dynamically generate the heat energy distribution strategy, including: obtaining the remaining capacity of the heat energy storage device, the predicted value of the heat energy demand, and the priority of the communication device; integrating the remaining capacity of the heat energy storage device, the predicted value of the heat energy demand, and the priority of the communication device, and inputting the temperature, humidity, and equipment load information of each area of the computer room into the deep Q-network reinforcement learning algorithm, and determining the target action through the Q value output by the network and the ε-greedy strategy; executing the target action, obtaining the new state and reward from the environment, storing the current state, action, reward, and new state as an experience tuple in the experience replay buffer, randomly sampling the experience tuple from the experience replay buffer, and updating the weights of the deep Q-network according to the reward and the target Q value, dynamically adjusting the policy parameters, evaluating and optimizing the policy, and generating the target heat energy distribution strategy.

[0067] Among them, the priority of the communication device can be a priority level set for it according to factors such as the importance of the communication device in the entire communication system and the service requirements. For example, for devices that undertake key services and have a greater impact on the communication service quality, a higher priority will be given; the experience tuple can be a quadruple composed of the current state, action, reward, and new state; the experience replay buffer can be a data structure for storing experience tuples.

[0068] It can be understood that the embodiments of the present application collect multi-dimensional data such as the remaining capacity of the heat energy storage device, the predicted value of the heat energy demand, and the priority of the communication device, use the deep Q-network reinforcement learning algorithm, combine the ε-greedy strategy to determine the action, and use the experience replay mechanism to continuously optimize the network weights and policy parameters, dynamically generate the target heat energy distribution strategy, realize the intelligent and precise allocation of the heat energy resources in the communication computer room, can effectively balance the temperature, humidity and equipment load in each area, ensure the stable operation of the communication device, and can also flexibly allocate the heat energy based on the prediction and priority, significantly improving the energy utilization efficiency.

[0069] Specifically, a large communication computer room contains multiple cabinet areas, each of which deploys communication devices of different importance levels, and the computer room is equipped with a phase change material thermal energy storage device. The remaining capacity of the thermal energy storage device is monitored in real time to be 60%. Through historical data and factors such as weather and business hours, it is predicted that within the next two hours, due to equipment operation and environmental changes, the thermal energy demand of the entire computer room will increase by 30%. At the same time, according to the service types carried by the equipment, the priority of the core switching equipment is determined to be high, and the priority of the ordinary data storage equipment is determined to be low. In addition, it is also obtained that in area A of the computer room, due to dense equipment, the temperature is relatively high and the load is relatively heavy, while in area B, there are fewer equipment, and both the temperature and the load are relatively low. These data are input into the deep Q-network reinforcement learning algorithm, and the Q values corresponding to different actions are output by the network. Combining with the ε-greedy strategy, the target action is determined to be to allocate part of the stored thermal energy to area A, and at the same time adjust the operating power of the air conditioner to reduce the heat generation in area A. After executing this action, the temperature in area A drops to a reasonable range, the equipment operates stably, and a positive reward is obtained; at the same time, the system records the current state, the executed action, the obtained reward, and the new environmental state, forms an experience tuple, and stores it in the experience replay buffer. Subsequently, experience tuples are randomly sampled from the buffer, and the weights of the deep Q-network are updated based on the reward and the target Q value, and the policy parameters are adjusted. After multiple iterations of optimization, a target thermal energy allocation strategy that can adapt to the complex environment of the computer room is finally generated, achieving a balance between high energy efficiency and stable equipment operation.

[0070] It should be noted that the ε-greedy strategy is a strategy used to balance exploration and exploitation in reinforcement learning. Among them, ε is a parameter between 0 and 1. When making a decision, an action is randomly selected with a probability of ε for exploration to discover new and potentially better actions; an action with the largest current Q value is selected with a probability of 1 - ε for exploitation, that is, the currently known optimal action is selected. By adjusting the value of ε, the balance degree between exploration and exploitation can be controlled. In the initial stage of the algorithm, a relatively large value of ε is usually set to explore the environment more fully; as learning progresses, the value of ε is gradually reduced to make more use of the optimal actions that have been learned.

[0071] In the embodiment of the present application, after generating the target thermal energy allocation strategy, it includes: obtaining the operating parameters of the thermal energy allocation; establishing a fault diagnosis model, inputting the operating parameters into the fault diagnosis model to determine whether there is a fault. If there is no fault, the operation is carried out; otherwise, the cause and location of the fault are automatically analyzed, and a detailed repair suggestion is generated; the thermal energy allocation strategy is automatically adjusted according to the modification suggestion.

[0072] It can be understood that in the embodiments of the present application, by obtaining the operating parameters of heat energy distribution, analyzing and judging the parameters using the established fault diagnosis model, when there is no fault, it operates normally, and when there is a fault, it automatically analyzes the cause and location of the fault, generates detailed repair suggestions, and automatically adjusts the heat energy distribution strategy according to the suggestions, realizing the fault self-diagnosis and strategy optimization of the heat energy distribution system, and ensuring the reliability and efficiency of operation.

[0073] Specifically, it is monitored that within a certain period, the pipeline flow rate in area A drops sharply, while the temperature rises abnormally, and the heat pump power fluctuates frequently. These parameters are input into the fault diagnosis model, and after the model analyzes and judges, it is found that there is a fault. The model automatically analyzes through algorithms and compares with the preset fault feature library that the pipeline in area A has a local blockage due to long-term use, resulting in poor flow, rising temperature, and affecting the normal operation of the heat pump at the same time. Immediately, detailed repair suggestions are generated, such as immediately closing the relevant pipeline valves in area A, arranging maintenance personnel to clean the blocked pipeline, and checking the circuit and compressor status of the heat pump. According to the repair suggestions, the system automatically adjusts the heat energy distribution strategy, temporarily redistributes the heat energy originally transported to area A to other areas, maintaining the overall heat energy supply and demand balance in the computer room until the fault is repaired, and then restoring the original distribution strategy.

[0074] According to the temperature energy-saving control method for communication computer rooms proposed in the embodiments of the present application, the temperature, humidity, and communication equipment load data of each area in the computer room are collected in real time through a multi-source sensor network, the heat dissipation risk level is quantified by a pre-trained equipment health assessment model, and the environmental adaptability dynamic diagnosis is realized by combining the equipment operation temperature-humidity-load correlation matrix, accurately identifying the areas where the environment exceeds the standard and triggering the local temperature control response mechanism to avoid the "one-size-fits-all" energy consumption waste of traditional unified regulation. Secondly, by integrating the refrigeration efficiency ratio calculation model and the dynamic weight multi-objective optimization algorithm, with the real-time electricity price, equipment remaining life, and performance attenuation coefficient as constraint variables, a candidate solution set that takes into account cost, efficiency, and reliability is generated through virtual simulation verification, and the optimal strategy is selected based on multi-dimensional comprehensive scoring, breaking the decision-making dilemma of the separation of cost and performance in the generation of traditional solutions. Introduce the collaborative utilization mechanism of air-conditioning waste heat recovery and cabinet heat dissipation, and combine the deep Q-network reinforcement learning algorithm to construct a dynamic heat energy distribution model, and realize the intelligent decision-making of heat energy storage, transfer, and reuse through reinforcement learning iterative optimization, converting low-efficiency waste heat into high-efficiency energy supply resources. Thus, the problems of energy waste, high cost, difficulty in accurately adapting to the heat dissipation requirements of each area, affecting the equipment life and stability, and increasing the risk of failure in the prior art are solved.

[0075] Next, a specific embodiment will be used to elaborate on the temperature energy-saving control method for communication computer rooms, and the specific content is as follows: A large Internet company has a large communication computer room in a first-tier city. The computer room covers an area of 3,000 square meters and has more than 500 cabinets deployed inside, carrying the company's core network services, data storage and processing operations. The equipment loads in this computer room vary significantly. The core server area undertakes a large number of user access and data computing tasks, while the load in the edge equipment area is relatively low.

[0076] High-precision temperature and humidity sensors and equipment load monitoring devices are installed in the computer room to collect temperature data, humidity data and communication equipment load data in each area in real time. As Figure 2 shown, at 9 am, the temperature in the core server area A was monitored to be 28°C, the humidity was 50%, and the power load reached 80%. At 9 am and 10 am, the temperature remained at 28°C; after adjustment, the temperature rose to 33°C; it is predicted that the temperature will further rise to 35°C from 14 to 16 pm; after implementing the target plan, the temperature dropped back to 28°C. The temperature in the edge equipment area B is 25°C, the humidity is 45%, and the power load is 30%.

[0077] At the same time, the system retrieves historical data for the past year and uses time series analysis algorithms to build a model. It is predicted that from 14 to 16 pm on the same day, due to a sharp increase in the number of user accesses, the communication equipment load in the core server area A will rise to 95%, and the temperature may rise to 35°C. Based on this prediction result, the system pre-adjusts the set value of the air-conditioning refrigeration temperature in this area from 26°C to 24°C in advance, increases the operating frequency of the ventilation equipment, and at the same time conducts predictive maintenance inspections on the key components of the refrigeration equipment to ensure the stable operation of the equipment under high load.

[0078] The collected temperature data, humidity data and communication equipment load data are input into a pre-trained equipment health assessment model. The model outputs that the cabinet heat dissipation risk level in the core server area A is a high risk (the equipment temperature has reached the threshold of 85% and the load continues to rise), and the edge equipment area B is a low risk. Combining with the preset communication equipment operating temperature-humidity-load correlation matrix, it is judged that the core server area A does not meet the operating requirements, and the local temperature control adjustment mechanism is triggered.

[0079] Based on the cabinet temperature data, the system adjusts the blade angle of the corresponding outlet floor in the core server area A from 30° to 60° to enhance air circulation. However, 15 minutes later, the cabinet temperature in this area continued to rise to 33°C, exceeding the preset condition of a 5°C difference from the average temperature of the computer room. Immediately, the local auxiliary refrigeration equipment - a small refrigeration fan - in this area is started to quickly reduce the cabinet temperature and ensure the normal operation of the equipment; while the edge equipment area B meets the operating requirements and enters the operating mode.

[0080] At 10 o'clock in the morning on the same day, outdoor meteorological data was obtained: temperature 23°C, humidity 55%, and wind speed 2.5 m / s. Since the outdoor temperature was between the first preset temperature (20°C) and the second preset temperature (30°C), the air-conditioning compressor in the computer room was turned off, the fan was kept running, and natural ventilation was started. Based on the temperature and humidity difference between indoor and outdoor, the real-time temperature and humidity data in the computer room, and the load condition of communication equipment, after calculation by the fuzzy control algorithm, the operation intensity of the ventilation equipment was adjusted to 70%, and the speed of the air-conditioning fan was adjusted to 60%, reducing energy consumption while ensuring the equipment operation environment.

[0081] After entering the operation mode and completing the local temperature control adjustment, the system calculated the refrigeration efficiency ratio based on the air volume of the ventilation equipment (currently 9000 cubic meters per hour), the power consumption of the air conditioner (the compressor was turned off, only the fan consumed power, with a power of 5 kilowatts), and the real-time electricity price (0.7 yuan per degree) data. Combining the remaining service life of the refrigeration equipment (expected to be 4 years) and the performance degradation coefficient (0.8), the following candidate control scheme set was generated through the dynamic weight multi-objective optimization algorithm, as Figure 3 shown: Scheme 1: Further reduce the power of the air-conditioning fan to 4 kilowatts and increase the air volume of the ventilation equipment to 11000 cubic meters per hour. It is expected that the refrigeration efficiency ratio will be increased to 1.4, but the temperature in some areas may rise.

[0082] Scheme 2: Maintain the current operation parameters of the equipment, regularly clean the filter of the ventilation equipment. It is expected that the refrigeration efficiency ratio will be increased to 1.3, and the remaining service life of the equipment will be extended by 1 month.

[0083] Scheme 3: Turn on the air-conditioning compressor and set the refrigeration temperature to 22°C to ensure the stability of the computer room temperature, but the power consumption of the air conditioner will increase to 30 kilowatts, and the refrigeration efficiency ratio will drop to 1.2.

[0084] Based on the digital twin technology, a virtual model for temperature and energy-saving control of the communication computer room was constructed, and the three candidate schemes were respectively input for operation. The simulation results showed that: after 3 hours of operation in Scheme 1, the temperature in the core server area exceeded 32°C, affecting the stability of the equipment; during the simulation period in Scheme 2, the computer room environment was stable, the energy consumption was low, and the equipment status was good; Scheme 3 had too high energy consumption and a significant increase in operating costs. Finally, Scheme 2 was determined as the target scheme because of the highest comprehensive score.

[0085] Based on the target scheme, the system stored the waste heat of the air conditioner and the heat dissipated by the cabinets. The current remaining capacity of the phase change material thermal energy storage equipment equipped in the computer room was 50%. By analyzing and predicting, it was determined that the thermal energy demand in the computer room would increase by 25% in the next 3 hours, and at the same time, the priority of the communication equipment in the core server area was determined to be high.

[0086] Integrate the temperature, humidity, and equipment load information of each area in the computer room from the above data and input it into the deep Q-network reinforcement learning algorithm. Combining with the ε-greedy strategy (the initial ε value is set to 0.8), determine the target action: allocate part of the stored thermal energy to the core server area and adjust the air conditioner operation mode to recover more waste heat. After executing this action, the temperature in the core server area drops to 28 °C, the equipment runs stably, and a positive reward is obtained. The system records the current state, action, reward, and new state, forms an experience tuple, and stores it in the experience replay buffer.

[0087] During subsequent operations, when the system obtains the operation parameters of thermal energy allocation, it is found that the pipeline flow rate in area C drops suddenly from the normal 50 cubic meters per hour to 10 cubic meters per hour, and the temperature rises rapidly from 30 °C to 38 °C. Input these parameters into the fault diagnosis model to determine that there is a blockage fault in the pipeline in area C. The model automatically analyzes that the blockage location is in the middle of the pipeline and generates repair suggestions: close the relevant valves in area C and arrange personnel to clean the pipeline. At the same time, according to the repair suggestions, the system automatically redistributes the thermal energy originally allocated to area C to other areas to maintain the overall thermal energy supply and demand balance in the computer room until the fault repair is completed and the original allocation strategy is restored.

[0088] In summary, the embodiment of this application takes the communication computer room of a large Internet company as the scenario. First, it uses sensors to collect real-time temperature, humidity, and equipment load data in multiple areas, combines historical data, uses time series algorithms to predict the change trend, and adjusts equipment parameters and maintenance in advance; then, based on the equipment health assessment model and the correlation matrix, it judges the environmental state, triggers local temperature control for non-compliant areas, and conducts hierarchical control according to the outdoor temperature for compliant areas; then, considering factors such as comprehensive energy consumption and equipment life, it generates candidate control plans, and selects the optimal target plan through verification by the digital twin virtual model; then, based on the target plan, it uses the deep Q-network reinforcement learning to achieve intelligent thermal energy allocation; finally, it monitors the operation parameters through the fault diagnosis model, automatically locates and processes thermal energy allocation faults and adjusts the strategy, and realizes energy-saving and efficient control of the computer room temperature throughout the process.

[0089] Next, refer to the accompanying drawings to describe the temperature energy-saving control system of the communication computer room proposed according to the embodiment of this application.

[0090] Figure 4 It is a block diagram of the temperature energy-saving control system of the communication computer room according to the embodiment of this application.

[0091] As Figure 4 shown, the temperature energy-saving control system 10 of the communication computer room includes: an acquisition module 100, a judgment module 200, a calculation module 300, and a generation module 400.

[0092] Among them, the acquisition module 100 is used to acquire the temperature data, humidity data, and communication device load data of multiple areas in the communication machine room; the judgment module 200 is used to input the temperature data, humidity data, and communication device load data into a pre-trained device health assessment model, output the heat dissipation risk level of the cabinet, and combine the preset communication device operating temperature-humidity-load association matrix to judge whether the environmental status of each area in the machine room meets the operating requirements. If the operating requirements are met, enter the operating mode; otherwise, trigger the local temperature control adjustment mechanism; the calculation module 300 is used to calculate the refrigeration efficiency ratio based on the ventilation equipment air volume, air conditioner power consumption, and real-time electricity price data after entering the operating mode or triggering the local temperature control adjustment mechanism, and combine the remaining service life and performance decay coefficient of the refrigeration equipment to generate a candidate control solution set through the dynamic weight multi-objective optimization algorithm, perform virtual verification on the candidate control solution set, and select the solution with the highest comprehensive score as the target solution; the generation module 400 is used to store heat based on the target solution according to the air conditioner waste heat and cabinet heat dissipation, and use the deep Q-network reinforcement learning algorithm to dynamically generate a heat energy distribution strategy.

[0093] It should be noted that the foregoing explanation of the embodiment of the temperature energy-saving control method for the communication machine room is also applicable to the temperature energy-saving control system of the communication machine room in this embodiment, and will not be elaborated here.

[0094] According to the temperature energy-saving control system of the communication machine room proposed in the embodiment of the present application, the temperature, humidity, and communication device load data of each area in the machine room are collected in real time through a multi-source sensor network, the heat dissipation risk level is quantified by a pre-trained device health assessment model, and the environmental adaptability dynamic diagnosis is realized by combining the device operating temperature-humidity-load association matrix, accurately identifying the areas where the environment exceeds the standard and triggering the local temperature control response mechanism to avoid the "one-size-fits-all" energy consumption waste of traditional unified regulation. Secondly, by integrating the refrigeration efficiency ratio calculation model and the dynamic weight multi-objective optimization algorithm, with the real-time electricity price, remaining equipment life, and performance decay coefficient as constraint variables, a candidate solution set that takes into account cost, efficiency, and reliability is generated through virtual simulation verification, and the optimal strategy is selected based on multi-dimensional comprehensive scoring, breaking the decision-making dilemma of the separation of cost and performance in traditional solution generation. Introduce the mechanism of coordinated utilization of air conditioner waste heat recovery and cabinet heat dissipation, combine the deep Q-network reinforcement learning algorithm to construct a dynamic heat energy distribution model, and realize the intelligent decision-making of heat energy storage, transfer, and reuse through reinforcement learning iteration optimization, converting low-efficiency waste heat into high-efficiency energy supply resources. Thus, the problems of energy waste, high cost, difficulty in accurately adapting to the heat dissipation requirements of each area, affecting the equipment life and stability, and increasing the failure risk in the prior art are solved.

[0095] Figure 5 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. The electronic device may include: A memory 501, a processor 502, and a computer program stored on the memory 501 and executable on the processor 502.

[0096] When the processor 502 executes the program, it implements the temperature energy-saving control method for the communication computer room provided in the above embodiments.

[0097] Furthermore, the electronic device further includes: A communication interface 503 for communication between the memory 501 and the processor 502.

[0098] The memory 501 is used to store a computer program executable on the processor 502.

[0099] The memory 501 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory.

[0100] If the memory 501, the processor 502, and the communication interface 503 are implemented independently, the communication interface 503, the memory 501, and the processor 502 can be interconnected through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 5 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0101] Optionally, in a specific implementation, if the memory 501, the processor 502, and the communication interface 503 are integrated on a chip, the memory 501, the processor 502, and the communication interface 503 can communicate with each other through an internal interface.

[0102] The processor 502 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0103] In the description of this specification, the descriptions with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0104] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of this application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0105] Any process or method description shown in a flowchart or described in other ways herein can be understood to represent a module, segment, or portion of code including one or N executable instructions for implementing a customized logical function or process, and the scope of the preferred embodiments of this application includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of this application belong.

[0106] It should be understood that each part of this application can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following well-known technologies in the art can be used: discrete logic circuits with logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits with suitable combinational logic gate circuits, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0107] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of implementing the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

Claims

1. A temperature energy-saving control method for a communication machine room, characterized in that It includes the following steps: Obtain the temperature data, humidity data, and communication equipment load data of multiple areas in the communication machine room; Input the temperature data, humidity data, and communication equipment load data into a pre-trained equipment health assessment model, output the heat dissipation risk level of the cabinet, and combine it with a preset communication equipment operating temperature-humidity-load correlation matrix to determine whether the environmental status of each area in the machine room meets the operating requirements. If it meets the operating requirements, enter the operating mode; otherwise, trigger the local temperature control adjustment mechanism; After entering the operating mode or triggering the local temperature control adjustment mechanism, calculate the refrigeration efficiency ratio based on the air volume of the ventilation equipment, air conditioning power consumption, and real-time electricity price data, and combine the remaining service life and performance attenuation coefficient of the refrigeration equipment. Generate a candidate control solution set through the dynamic weight multi-objective optimization algorithm, perform virtual verification on the candidate control solution set, and select the solution with the highest comprehensive score as the target solution; Based on the target solution, store heat according to the air conditioning waste heat and cabinet heat dissipation, and use the deep Q-network reinforcement learning algorithm to dynamically generate a heat energy distribution strategy.

2. The temperature energy-saving control method for a communication machine room according to claim 1, wherein If the operating requirements are met, enter the operating mode, including: Obtain outdoor meteorological data, where the outdoor meteorological data includes temperature, humidity, wind speed, etc.; If the outdoor temperature is less than or equal to the first preset temperature, turn off the air conditioner and start natural ventilation. At the same time, dynamically optimize the opening degree of the ventilation opening and the angle of the deflector according to the wind speed and wind direction data through the fluid dynamics simulation model; If the outdoor temperature is greater than the first preset temperature and less than or equal to the second preset temperature, turn off the air conditioner and start natural ventilation. Dynamically adjust the operating intensity of the air conditioner and ventilation equipment based on the indoor and outdoor temperature and humidity difference, the real-time temperature and humidity data in the machine room, and the communication equipment load; If the outdoor temperature is greater than the second preset temperature, run the air conditioner at full power in advance, and calculate the optimal air outlet angle combination of the air conditioner group according to the hot spot distribution in the machine room through the ant colony optimization algorithm.

3. The temperature energy-saving control method for a communication machine room according to claim 1, characterized in that, Trigger the local temperature control adjustment mechanism, including: adjust the blade angle of the air outlet floor according to the cabinet temperature data, and monitor the change trend of the cabinet temperature data in real time. When it is found that the cabinet temperature continues to rise and reaches the preset condition, start the local auxiliary refrigeration equipment.

4. The temperature energy-saving control method for a communication machine room according to claim 3, characterized in that, The preset condition is that the difference between the cabinet temperature and the average temperature of the machine room exceeds 5°C, or the cabinet temperature reaches the preset temperature threshold.

5. The temperature energy-saving control method for a communication machine room according to claim 1, characterized in that, Use the deep Q-network reinforcement learning algorithm to dynamically generate a heat energy distribution strategy, including: Obtain the remaining capacity of the heat energy storage device, the predicted value of heat energy demand, and the priority of communication equipment; Integrate the remaining capacity of the heat energy storage device, the predicted value of heat energy demand, and the priority of communication equipment, and input the temperature and humidity and equipment load information of each area in the machine room into the deep Q-network reinforcement learning algorithm to determine the target action through the Q value output by the network and the ε-greedy strategy; Execute the target action, obtain the new state and reward from the environment, store the current state, action, reward, and new state as an experience tuple in the experience replay buffer, randomly sample experience tuples from the experience replay buffer, update the weights of the deep Q-network according to the reward and the target Q-value, dynamically adjust the policy parameters, evaluate and optimize the policy, and generate the target thermal energy allocation policy.

6. The temperature energy-saving control method for a communication machine room according to claim 5, characterized in that, After generating the target thermal energy allocation policy, it includes: Obtain the operating parameters of the thermal energy allocation. Establish a fault diagnosis model, input the operating parameters into the fault diagnosis model, and determine whether there is a fault. If there is no fault, proceed with the operation; otherwise, automatically analyze the cause and location of the fault and generate detailed repair suggestions. Automatically adjust the thermal energy allocation policy according to the modification suggestions.

7. The temperature energy-saving control method for a communication machine room according to claim 1, characterized in that, Perform virtual verification on the candidate control scheme set, and select the scheme with the highest comprehensive score as the target scheme, including: Construct a virtual model for temperature energy conservation control of the communication room based on digital twin technology. Input the target scheme into the virtual model for operation, and obtain the energy conservation effect of the scheme, the operation stability of the equipment, and the impact on the operating environment of the communication equipment. If the simulation results do not meet the expected standards, iterate and optimize the target scheme.

8. The temperature energy-saving control method for a communication machine room according to claim 1, characterized in that After obtaining the temperature data, humidity data, and communication equipment load data of multiple areas in the communication room, it further includes: Obtain historical data. Use time series analysis algorithms to model the historical data, and predict the temperature and humidity change trends of each area in the computer room and the change of the communication equipment load. According to the prediction results, adjust the operating mode or the parameters of the local temperature control adjustment mechanism in advance, and perform predictive maintenance on the refrigeration equipment and ventilation equipment.

9. The temperature energy-saving control method for a communication machine room according to claim 1, characterized in that The formula of the equipment health assessment model is: Among them, is the health assessment value of the i-th cabinet, is the current temperature of the i-th cabinet, is the current humidity of the i-th cabinet, is the current load value of the i-th cabinet, is the maximum allowable threshold of temperature, is the maximum allowable threshold of humidity, is the maximum allowable threshold value of the load value, is the temperature weight coefficient, is the humidity weight coefficient, is the load value weight coefficient.

10. A temperature energy-saving control system for a communication machine room, characterized in that It includes: An acquisition module for acquiring the temperature data, humidity data, and communication equipment load data of multiple areas in the communication room. A judgment module for inputting the temperature data, humidity data, and communication equipment load data into a pre-trained equipment health assessment model, outputting the heat dissipation risk level of the cabinet, and combining a preset communication equipment operating temperature-humidity-load correlation matrix to judge whether the environmental status of each area in the computer room meets the operating requirements. If the operating requirements are met, enter the operating mode; otherwise, trigger the local temperature control adjustment mechanism. A calculation module for calculating the refrigeration efficiency ratio based on the ventilation equipment air volume, air conditioner power consumption, and real-time electricity price data after entering the operating mode or triggering the local temperature control adjustment mechanism, and generating a candidate control scheme set through a dynamic weight multi-objective optimization algorithm in combination with the remaining service life and performance degradation coefficient of the refrigeration equipment, and performing virtual verification on the candidate control scheme set, and selecting the scheme with the highest comprehensive score as the target scheme. A generation module for storing heat based on the waste heat of the air conditioner and the heat dissipation of the cabinet according to the target scheme, and dynamically generating a thermal energy allocation policy using the deep Q-network reinforcement learning algorithm.

Images