Intelligent condensation regulation system and method for two-phase flow jet atomization

Abstract

The application discloses a two-phase flow jet flow atomization coordinated intelligent condensation regulation and control system and method, relates to the technical field of condensation regulation and control, and comprises the following steps: S1, real-time acquisition of coordinated condensation regulation and control data, data preprocessing; S2, real-time discrimination of condensation working conditions, switching of jet flow atomization coordinated condensation modes according to the condensation working conditions; S3, entering the jet flow atomization coordinated condensation mode, evaluation of the intensity of atomization spraying, and hierarchical regulation and control of jet flow atomization coordinated condensation; S4, real-time evaluation of liquid level and energy efficiency comprehensive safety in the jet flow atomization coordinated condensation regulation and control process, adjustment of liquid discharge and condensation operation modes based on the liquid level and energy efficiency comprehensive safety evaluation results. The problems that, for high heat flow density data centers, the existing two-phase immersion liquid cooling system condensation link mainly adopts passive condensation, lacks active intelligent regulation and control, and steam accumulation, cavity pressure fluctuation and condensation lagging are prone to occur under high load, thereby affecting heat dissipation stability and efficiency are solved.

Description

Technical Field

[0001] This invention relates to the field of condensation control technology, specifically to an intelligent condensation control system and method for coordinated two-phase jet atomization. Background Technology

[0002] With the rapid development of digital technology, data centers, as the core infrastructure for information processing and storage, are experiencing continuously increasing equipment power density, making heat dissipation a key factor in ensuring stable system operation. Traditional air-cooling technology, limited by the heat transfer performance of the heat dissipation medium, is no longer sufficient to meet the heat dissipation needs of high heat flux density equipment. Liquid cooling technology, with its higher heat exchange efficiency, is gradually becoming the mainstream choice. Currently, two-phase immersion liquid cooling technology is widely used in high heat flux density server cooling scenarios. It absorbs heat through a phase change in the working medium, and then converts the vapor into liquid through a condensation device to achieve circulating heat dissipation, demonstrating strong heat dissipation capabilities.

[0003] For example, invention patent CN118170233A discloses a two-phase liquid cooling system for servers. Multiple two-phase liquid cooling subsystems correspond one-to-one with the upper and lower layers of servers within a server rack. Each two-phase liquid cooling subsystem includes a liquid cooling plate, a condenser, gas pipes, and liquid pipes. The liquid cooling plate is horizontally attached to the server, and the condenser is positioned higher than the liquid cooling plate. Both the gas pipes and liquid pipes are gravity-type heat pipes and are connected between the liquid cooling plate and the condenser to form a self-driven refrigerant circulation loop. A multi-source heat exchanger is used to cool the condensers of the multiple two-phase liquid cooling subsystems. This invention offers efficient, reliable, and safe heat dissipation. The system is simple, low-cost, and easy to maintain, making it suitable for high-density server rack cooling.

[0004] For example, invention patent CN113778205B discloses a multi-unit forced two-phase immersion liquid cooling system and its control method, comprising: a forced two-phase immersion liquid cooling device, a gravity heat pipe heat exchange device, a natural cooling device, and a valve control system; the forced two-phase immersion liquid cooling device, the gravity heat pipe heat exchange device, the valve control system, and the natural cooling device are connected sequentially; the natural cooling device includes an indirect cooling device and an evaporative cooling device; the indirect cooling device and the evaporative cooling device are respectively connected to the valve control system. This invention uses natural cooling, which can effectively save energy.

[0005] However, for data centers with high heat flux, the condensation stage of two-phase immersion liquid cooling systems is still mainly passive condensation, lacking active and intelligent control over vapor migration and condensation intensity. This leads to vapor accumulation, cavity pressure fluctuations, and condensation lag under high loads, making it difficult to ensure stable and efficient heat dissipation.

[0006] Therefore, in order to address the above problems, there is an urgent need for an intelligent condensation control system and method that coordinates two-phase jet atomization. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides an intelligent condensation control system and method for coordinated two-phase jet atomization. This solves the problem that existing two-phase immersion liquid cooling systems for high heat flux density data centers rely primarily on passive condensation, lacking active intelligent control. Consequently, these systems are prone to vapor accumulation, cavity pressure fluctuations, and condensation lag under high loads, affecting heat dissipation stability and efficiency.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: an intelligent condensation control method for two-phase flow jet atomization synergy, comprising the following steps: S1, real-time acquisition of synergistic condensation control data, and data preprocessing of the synergistic condensation control data; S2, real-time determination of condensation conditions based on the preprocessed synergistic condensation control data, and switching of jet atomization synergistic condensation mode according to the condensation conditions; S3, entering the jet atomization synergistic condensation mode, and evaluating the intensity of atomization spray based on the preprocessed synergistic condensation control data, and performing graded control of jet atomization synergistic condensation based on the atomization spray intensity evaluation results; S4, during the jet atomization synergistic condensation control process, real-time evaluation of liquid level and energy efficiency safety based on the preprocessed synergistic condensation control data, and adjusting the drainage and condensation operation modes based on the liquid level and energy efficiency safety evaluation results.

[0011] Furthermore, the real-time acquisition and preprocessing of collaborative condensation control data involves the following steps: Real-time acquisition of collaborative condensation control data is achieved using various types of sensors deployed at key nodes in the steam chamber, liquid cooling pipelines, and server, including steam chamber pressure sensors, liquid level sensors, nozzle flow sensors, and intelligent power distribution units. This data includes: steam chamber pressure, tank liquid level, server heat load, total energy consumption of the condensation control equipment, overall total operating energy consumption, nozzle flow rate, maximum rated nozzle flow rate, and latent heat of vaporization of the coolant. The collaborative condensation control data is then denoised and smoothed using moving average and median filtering algorithms, eliminating sudden extreme values ​​and invalid data, and unifying timestamp calibration and data alignment. For signals with missing or abnormal values, integrity correction is performed using short-time historical interpolation, data source redundancy switching, and physical model verification methods. Simultaneously, the collaborative condensation control data is standardized and normalized, and consistency logic verification is performed. Finally, a collaborative condensation control database is established to store the collaborative condensation control data.

[0012] Furthermore, based on the preprocessed collaborative condensation control data, the specific process for real-time determination of condensation conditions is as follows: Based on a sliding time window, historical steam chamber pressures are obtained, the historical steam chamber pressure data are sorted in ascending order, and the P95 quantile is selected using the quantile method to obtain the steam chamber pressure benchmark value. The current steam chamber pressure is divided by the steam chamber pressure benchmark value to obtain the steam chamber pressure margin ratio. The server heat load and the latent heat of vaporization constant of the coolant are obtained, and the ratio of the current server heat load to the latent heat of vaporization constant of the coolant is calculated to obtain the current steam yield. Based on a sliding time window, historical server heat loads are obtained, the historical server heat load data are sorted in ascending order, and the P95 quantile is selected using the quantile method to obtain the typical load. The ratio of the typical load to the latent heat of vaporization constant of the coolant is calculated to obtain the benchmark steam yield. The current steam yield is divided by the benchmark steam yield to obtain the active steam yield value. The active steam yield value is added to a constant and a natural logarithmic operation is performed to obtain the steam load response value. The steam chamber pressure margin ratio is multiplied by the steam load response value to obtain the condensation condition determination value.

[0013] Furthermore, the specific process of switching the jet atomization-coordinated condensation mode according to the condensation condition is as follows: The condensation condition discrimination value is calculated in real time, and the PID controller performs graded switching of condensation regulation based on the discrimination value: When the condensation condition discrimination value is less than the condensation threshold, it is determined to be in normal operating condition, and the jet condensation mode is activated: only the Venturi tube on the high-efficiency condenser coil is driven for jetting, without the need to start spray atomization, and the pump is maintained in normal operation; at the same time, the jet pump frequency is optimized based on historical coordinated condensation regulation data; when the condensation condition discrimination value is greater than or equal to the condensation threshold, it is determined that condensation is insufficient, and generation... Upon receiving the enhanced condensation command, the atomizing spray device on the high-efficiency condensing coil is activated, entering the jet atomization-coordinated condensation mode. After entering enhanced condensation, the condensation condition judgment value is continuously monitored. If the condensation condition judgment value falls back below the condensation threshold, the electronically controlled valve is closed to shut off the atomizing nozzle, resuming jet condensation. The condensation condition judgment value is written into the coordinated condensation control database, recording the condensation mode switching and corresponding coordinated condensation control data throughout the process. Based on real-time feedback and historical response results, reinforcement learning and Bayesian optimization algorithms are used to optimize the steam chamber pressure benchmark, typical load, and condensation threshold.

[0014] Further, the specific process of entering the jet atomization synergistic condensation mode and evaluating the intensity of atomization spraying based on the pre-processed synergistic condensation control data is as follows: receiving the enhanced condensation command and controlling the atomization spraying equipment to perform atomization spraying control; obtaining the current steam chamber pressure and the steam chamber pressure reference value, dividing the difference between the current steam chamber pressure and the steam chamber pressure reference value by the steam chamber pressure reference value to obtain the pressure over-limit amplitude ratio, adding the pressure over-limit amplitude ratio to a constant, and performing natural logarithmic calculation to obtain the pressure over-limit response value; obtaining the current steam yield and the nozzle's rated maximum flow rate, calculating the ratio of the current steam yield to the nozzle's rated maximum flow rate to obtain the steam load capacity ratio; multiplying the pressure over-limit response value by the steam load capacity ratio to obtain the atomization spraying synergistic control intensity value.

[0015] Furthermore, based on the intensity assessment results of the atomized spray, the specific process for graded control of jet atomization synergistic condensation is as follows: the intensity value of atomized spray synergistic control is written into the synergistic condensation control database, and the intensity value of atomized spray synergistic control is compared with the multi-level intensity thresholds T1 and T2 in real time; when the intensity value of atomized spray synergistic control is less than T1, it enters the synergistic condensation energy-saving standby state; when the intensity value of atomized spray synergistic control is greater than or equal to T1 and less than T2, it enters the synergistic condensation normal adjustment state; when the intensity value of atomized spray synergistic control is greater than or equal to T2, it enters the synergistic condensation enhanced protection state; at the same time, corresponding condensation control measures are taken for different synergistic condensation states.

[0016] Furthermore, the specific process for implementing corresponding condensation control measures for different synergistic condensation states is as follows: When entering the synergistic condensation energy-saving standby state, maintain the minimum spray and jet flow rate and reduce the pump frequency; when the time when the atomized spray synergistic control intensity value is less than T1 exceeds the maximum allowable threshold, the pump operates intermittently, and the nozzles only perform periodic self-cleaning pulses; when entering the synergistic condensation normal adjustment state, the jet intensity is increased by linearly mapping the atomized spray synergistic control intensity value to the pump speed, and the spray flow rate is increased by controlling the nozzles through electronically controlled valves; simultaneously, the spray is adjusted slowly in advance according to the server heat load change rate. The system analyzes the intensity of atomized spraying and uses an autoregressive moving average algorithm to predict future server heat load trends, optimizing the routine adjustment rhythm of coordinated condensation. When entering the enhanced coordinated condensation protection state, the nozzles are fully opened at the rated maximum flow rate, and the maximum jet flow rate is activated, while the backup drain valve is activated to prevent sudden liquid level rises. When the time when the coordinated control intensity value of atomized spraying is greater than or equal to T2 exceeds the maximum allowable threshold, an inspection work order is generated and an alarm push is initiated. The coordinated control intensity value of atomized spraying, nozzle flow rate, and the operating status of the radiator and condensation control equipment are recorded each time, and the coordinated control intensity value of atomized spraying and multi-level intensity thresholds are continuously optimized using a Bayesian optimization algorithm.

[0017] Furthermore, in the process of jet atomization synergistic condensation control, the specific process of real-time evaluation of the comprehensive safety of liquid level and energy efficiency based on the pre-processed synergistic condensation control data is as follows: Real-time reception of the atomization spray synergistic control intensity value, nozzle flow rate, and operating status of the radiator and condensation control equipment; during the jet atomization synergistic condensation process, simultaneously acquiring the tank liquid level, the total energy consumption of the condensation control equipment, and the overall total operating energy consumption; based on a sliding time window, acquiring historical tank liquid level data and sorting it in ascending order; selecting the P5 quantile of the tank liquid level within the window as the lower limit of liquid level safety, and selecting the P95 quantile of the tank liquid level within the selected window as the upper limit of liquid level safety; dividing the difference between the current tank liquid level and the lower limit of liquid level safety by the difference between the upper and lower limits of liquid level safety to obtain the relative safety margin value of liquid level; dividing the current total energy consumption of the condensation control equipment by the current total operating energy consumption to obtain the condensation energy consumption ratio; multiplying the condensation energy consumption ratio by the energy efficiency weighting factor and adding it to the relative safety margin value of liquid level to obtain the liquid level energy efficiency safety assessment value.

[0018] Furthermore, based on the comprehensive safety assessment results of liquid level and energy efficiency, the specific process for adjusting the drainage and condensation operation modes is as follows: The liquid level energy efficiency safety assessment value is compared with the energy efficiency threshold in real time. When the liquid level energy efficiency safety assessment value is less than the energy efficiency safety threshold, it is determined that both the liquid level and energy consumption are in a safe state. The corresponding jet atomization-coordinated condensation mode is maintained according to the atomization spray coordinated control intensity value. Simultaneously, the liquid level energy efficiency safety assessment value is continuously monitored for early warning. When the liquid level energy efficiency safety assessment value is greater than or equal to the energy efficiency safety threshold, it is determined that there is a risk to the liquid level and condensation energy consumption. The drainage valve is immediately activated to drain excess condensate back to the coolant tank, forcibly reducing the nozzle opening and pump speed, and decreasing the nozzle flow rate and jet intensity. If the time for which the liquid level energy efficiency safety assessment value is greater than or equal to the energy efficiency safety threshold exceeds the maximum allowable threshold, the drainage intensity is increased and the atomization and jet power are downgraded. All energy efficiency safety assessment values ​​and control actions are written into the coordinated condensation control database, and the energy efficiency safety threshold is optimized and corrected based on historical feedback information.

[0019] The second aspect of this invention provides an intelligent condensation control system for two-phase flow jet atomization synergy, comprising: a condensation data acquisition and preprocessing module for real-time acquisition of synergistic condensation control data and preprocessing the synergistic condensation control data; a condensation condition discrimination and condensation switching module for real-time discrimination of condensation conditions based on the preprocessed synergistic condensation control data and switching the jet atomization synergistic condensation mode according to the condensation conditions; an atomization spray enhanced condensation control module for entering the jet atomization synergistic condensation mode and evaluating the intensity of atomization spray based on the preprocessed synergistic condensation control data, and performing graded control of jet atomization synergistic condensation based on the intensity evaluation results; and a liquid level and energy efficiency safety closed-loop control module for real-time evaluation of liquid level and energy efficiency safety based on the preprocessed synergistic condensation control data during the jet atomization synergistic condensation control process, and adjusting the drainage and condensation operation modes based on the liquid level and energy efficiency safety evaluation results.

[0020] Beneficial effects

[0021] The present invention has the following beneficial effects:

[0022] (1) Based on the sliding window quantile method and adaptive criteria, the present invention dynamically extracts the pressure, load and liquid level operating condition benchmarks, realizes the automatic optimization of each core parameter of condensation control, and eliminates the need for manual setting of fixed thresholds, thus greatly improving adaptability and accuracy.

[0023] (2) By constructing a condensation condition discrimination value that combines the pressure margin ratio and the steam yield response factor, the present invention can determine the matching status of condensation capacity and load in real time, intelligently switch the jet and atomization enhancement modes, and improve the intelligent response capability and energy efficiency of condensation regulation.

[0024] (3) This invention achieves continuous adaptive adjustment of core control parameters such as spray flow rate and pump speed by constructing a graded and linear mapping algorithm for coordinated control of atomized spraying, and introduces a short-cycle load prediction and graded inspection linkage mechanism to improve control accuracy and fault prevention level.

[0025] (4) This invention constructs a joint safety assessment criterion for liquid level and energy efficiency, and weights and integrates the safety margin of liquid level with the proportion of energy consumption to achieve real-time closed-loop control of liquid level and energy consumption with dual objectives. It also continuously optimizes the safety threshold through historical feedback and self-learning to enhance safety and operational economy.

[0026] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0027] Figure 1 A flowchart of an intelligent condensation control method for coordinated atomization of two-phase jets;

[0028] Figure 2 A block diagram of an intelligent condensation control system for coordinated atomization of two-phase jets;

[0029] Figure 3 Schematic diagram of an intelligent condensation control system for coordinated atomization of two-phase jets;

[0030] Figure 4 This is a schematic diagram of a high-efficiency condenser coil structure;

[0031] Figure 5 This is a schematic diagram of the PID control condensation condition discrimination process;

[0032] Figure 6 This is a trend chart of liquid level energy efficiency safety assessment values ​​and assessment indicators.

[0033] In the diagram, 1. High-efficiency condenser coil; 2. Server; 3. Pump; 4. Coolant tank; 5. Radiator; 6. Electrically controlled valve; 7. Nozzle; 8. Steam chamber pressure sensor; 9. PID controller. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. As those skilled in the art will understand, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] Please see Figures 1-6 This invention provides a technical solution: an intelligent condensation control system and method for coordinated two-phase jet atomization, such as... Figure 1 As shown, the process includes the following steps: S1, real-time acquisition of collaborative condensation control data, and data preprocessing of the collaborative condensation control data; S2, based on the preprocessed collaborative condensation control data, real-time determination of condensation conditions, and switching of jet atomization collaborative condensation mode according to the condensation conditions; S3, entering the jet atomization collaborative condensation mode, and evaluating the intensity of atomization spray based on the preprocessed collaborative condensation control data, and performing graded control of jet atomization collaborative condensation based on the intensity evaluation results; S4, during the jet atomization collaborative condensation control process, real-time evaluation of liquid level and energy efficiency safety based on the preprocessed collaborative condensation control data, and adjusting the drainage and condensation operation modes based on the comprehensive safety evaluation results of liquid level and energy efficiency.

[0036] Specifically, the real-time acquisition and preprocessing of collaborative condensation control data involves the following steps: Real-time acquisition of collaborative condensation control data is achieved using various types of sensors deployed at multiple key nodes in the steam chamber, liquid cooling pipeline, and server terminals. These sensors include a steam chamber pressure sensor 8, a liquid level sensor, a nozzle flow sensor, and an intelligent power distribution unit. The collaborative condensation control data includes: steam chamber pressure, tank liquid level, server heat load, total energy consumption of the condensation control equipment, total overall operating energy consumption, nozzle flow rate, maximum rated nozzle flow rate, and latent heat of vaporization of the coolant. The total energy consumption of the condensation control equipment includes that of pump 3, nozzle 7, and electricity. The total electrical energy consumed during operation of equipment directly related to condensation enhancement and flow control, including control valve 6, high-efficiency condensing coil 1, drain valve, and PID controller 9; the total overall operating energy consumption is the sum of the power metering of all electrical equipment during operation, including not only the energy consumption of pump 3, nozzle 7, electrically controlled valve 6, high-efficiency condensing coil 1, drain valve, and PID controller 9 in the condensation control link, but also the coolant tank, radiator 5, temperature control module, energy consumption acquisition module, control unit, sensor network, and all other operating-related electrical equipment; the latent heat of vaporization constant of coolant is the heat absorbed per unit mass when coolant vaporizes under standard conditions. The collaborative condensation control data is denoised and smoothed using moving average and median filtering algorithms to remove sudden extreme values ​​and invalid data. Data alignment is achieved through unified timestamp calibration using the system clock, ensuring that multi-sensor data accurately correspond to the same sampling time. Data alignment ensures that all variables can be matched in chronological order. For signals with missing or abnormal data, integrity correction is performed using short-time historical interpolation, data source redundancy switching, and physical model verification. Short-time historical interpolation uses data from adjacent time points to predict missing points; data source redundancy switching refers to switching to a backup sensor; and physical model verification uses physical laws and thermodynamic equilibrium formulas to detect and correct abnormal data. Simultaneously, the collaborative condensation control data is standardized and normalized, and consistency logic checks are performed, including but not limited to physical quantity constraint checks, extreme value alarms, and mutual exclusion state detection, ensuring reasonable and conflict-free relationships between data. A collaborative condensation control database is established to store the collaborative condensation control data.

[0037] like Figure 3The diagram shows the principle of an intelligent condensation control system with two-phase flow jet atomization coordination. Server 2 is completely immersed in the liquid cooling medium. The equipment generates heat during operation, which is absorbed by the liquid cooling medium and triggers local evaporation, forming saturated steam. Pressure changes within the steam chamber are collected in real-time by a steam chamber pressure sensor 8. A high-efficiency condensing coil 1 is installed at the top of the steam chamber, serving as a condensation component. Coolant circulates within the coil, directly condensing the steam within the chamber. Nozzles 7 are integrated on the coil for atomization spraying to improve condensation efficiency. Pump 3 drives the coolant to circulate in the condensation loop, ensuring the jet velocity meets the condensation heat exchange requirements and providing power for the atomization spray branch flow. Coolant tank 4 serves as a storage and buffer node for the liquid cooling medium, stabilizing the liquid level, receiving returned condensate, and ensuring continuous liquid supply. In conjunction with a liquid level sensor, it enables safe liquid level monitoring and automatic drainage management to prevent overflow and supply interruption. Radiator 5 provides secondary cooling to the circulating coolant, dissipating the absorbed heat to the external environment, preventing a continuous rise in coolant temperature, and maintaining the thermal balance of the entire circulation loop. An electrically controlled valve 6 is installed in the fluid passage between the nozzle 7 and the high-efficiency condensing coil 1. It can adjust the opening degree according to operating conditions and criteria, achieving dynamic and precise control of the atomized spray flow rate and switching modes. The nozzle 7 is used to atomize the coolant and spray it into the steam chamber under high load or increased pressure, achieving enhanced condensation. Combined with the venturi tube, it forms a jet atomization synergistic condensation, efficiently cooling the steam. The PID controller 9, as the core intelligent control unit, integrates control algorithms, hierarchical criteria, and historical feedback, outputting real-time control commands for condensation mode switching, atomized spray opening degree, and pump speed adjustment, achieving closed-loop intelligent condensation control throughout the entire process.

[0038] like Figure 4 The diagram shows a schematic of a high-efficiency condensing coil. The high-efficiency condensing coil 1 consists of multiple coils arranged in parallel, serving as the main heat exchange and condensation unit for liquid cooling. Driven by pump 3, the coolant is evenly distributed into each coil from one end, absorbing heat from the steam chamber and then flowing back to the coolant tank 4, achieving efficient heat exchange and steam condensation. The coil body uses high thermal conductivity materials and an optimized layout to maximize condensation efficiency. Multiple Venturi tubes are evenly arranged at the top of each condensing coil. The coolant flows at high speed through these Venturi tubes, forming a jet that directly enters the steam chamber, actively disturbing and guiding the steam flow field within the chamber. The local negative pressure and high velocity generated by the Venturi effect not only accelerate the migration of steam to the condensation surface but also inhibit steam layer accumulation, enhancing the local heat exchange effect. The array design of multiple Venturi tubes ensures uniform disturbance of the flow field inside the steam chamber, enhancing condensation without dead zones and avoiding local overheating and uneven condensation. Working in conjunction with the nozzle, it enables the coordinated switching and intelligent graded control of jet condensation and atomization condensation. Under conditions of high heat flux density and excessive pressure, the jet from the Venturi array can couple with the atomized microdroplets from the top nozzle 7, further improving condensation efficiency.

[0039] This implementation plan achieves comprehensive real-time acquisition of multi-dimensional coordinated condensation control data, including pressure, liquid level, flow rate, and energy consumption. Combined with multi-level preprocessing techniques such as moving average, median filtering, and timestamp calibration, it effectively improves data accuracy and robustness. Historical interpolation, redundancy switching, and physical model verification ensure the integrity and continuity of the data chain. Simultaneously, the standardization, normalization, and consistency logic verification process enables seamless integration of various data types with subsequent judgment algorithms and intelligent control strategies, providing a high-quality data foundation and solid guarantee for intelligent condensation control, anomaly response, and safe operation.

[0040] Specifically, the real-time determination of condensation conditions based on preprocessed collaborative condensation control data is as follows: Based on a sliding time window, historical steam chamber pressures are acquired, and the historical steam chamber pressure data are sorted in ascending order. The P95 quantile is selected using the quantile method to obtain the steam chamber pressure benchmark value. The sliding time window refers to using historical data of a fixed duration as the statistical basis. The P95 quantile in the quantile method represents the pressure value greater than or equal to 95% of the samples in all collected data, filtering out extreme abnormal fluctuations and making the benchmark value more robust and adaptive. The current steam chamber pressure is divided by the steam chamber pressure benchmark value to obtain the steam chamber pressure margin ratio. The steam chamber pressure margin ratio measures the relative degree of the current pressure to the benchmark pressure, reflecting the physical index of the current load and safety margin. The server heat load and the latent heat of vaporization constant of the coolant are acquired, and the ratio of the current server heat load to the latent heat of vaporization constant of the coolant is calculated to obtain the current steam yield. Based on the sliding time window, the historical steam chamber pressures are obtained... Historical server thermal loads are collected and sorted in ascending order. The P95 quantile is selected using the quantile method to obtain the typical load. The typical load is used as a benchmark for comparison, which helps to dynamically reflect long-term high load characteristics and enhance the adaptability of the criterion. The ratio of the typical load to the latent heat of vaporization of the coolant is calculated to obtain the benchmark steam yield. The current steam yield is divided by the benchmark steam yield to obtain the steam yield activity value. The steam yield activity value reflects the activity of the current yield relative to historical high load conditions and is a key variable for measuring pressure response sensitivity. The steam yield activity value is added to a constant and the natural logarithm is performed to obtain the steam load response value. The natural logarithmic transformation can compress large-range variables, enhance the response sensitivity in the high load range, and suppress the influence of extreme anomalies on the criterion. The steam chamber pressure margin ratio is multiplied by the steam load response value to obtain the condensing condition discrimination value. The condensing condition discrimination value is the core decision basis for subsequent condensing mode switching and graded control.

[0041] The specific formula for the condensation condition discrimination value is as follows:

[0042] ;

[0043] In the formula, This represents the condensation condition judgment value, which comprehensively evaluates the combined impact of pressure risk and heat load surge. As a trigger condition for control logic, it not only measures the safety margin of condensation pressure but also sensitively captures the change in operating conditions due to sudden load increases. It realizes intelligent, real-time, and hierarchical judgment of condensation operation status, providing a quantitative basis for switching condensation strategies and ensuring the thermal safety of data centers. It represents the current steam chamber pressure, reflects the relationship between instantaneous load and condensation capacity, and is a key indicator of operating conditions; This represents the reference value for steam chamber pressure, used to dynamically adapt to operating conditions, serving as a benchmark for the current pressure and eliminating the influence of occasional anomalies. The steam chamber pressure margin ratio measures the relative level of the current pressure to the pressure under typical operating conditions. A steam chamber pressure margin ratio greater than 1 indicates that the pressure is too high, which may indicate insufficient condensation or excessive load. A steam chamber pressure margin ratio less than 1 indicates that the pressure safety margin is large and the operating condition is healthy. It indicates the current steam yield, reflecting in real time the actual requirements of the current heat load on the condensation process; This represents the baseline steam yield, indicating the normal steam load under conventional operating conditions. It represents the steam load response value, using a logarithmic amplifier, reflecting the degree of surge when the current steam load exceeds the reference load. It is not sensitive to normal small fluctuations, but responds quickly to overload conditions, preventing extreme value runaway.

[0044] This implementation scheme achieves adaptive dynamic extraction of steam chamber pressure and server heat load baseline values, effectively filtering extreme abnormal fluctuations and ensuring the robustness and representativeness of the condensing condition judgment criteria. A judgment logic coupling the steam chamber pressure margin ratio and steam yield activity value is constructed, and a natural logarithmic transformation is used to improve the response sensitivity in high-load ranges, enabling the condensing condition judgment value to accurately reflect the current pressure and load status and safety margin. This data-driven judgment method provides a scientific, accurate, and real-time decision-making basis for condensing mode switching and graded control, improving adaptability and safety efficiency.

[0045] Specifically, the process of switching the jet atomization-coordinated condensation mode according to the condensation condition is as follows: The condensation condition discrimination value is calculated in real time, and the PID controller 9 performs graded switching of condensation regulation based on the condensation condition discrimination value: When the condensation condition discrimination value is less than the condensation threshold, it is determined to be in normal operating condition, and the jet condensation mode is activated: Only the Venturi tube on the high-efficiency condensing coil 1 is driven to perform jetting. The Venturi tube structure is integrated at the top of the high-efficiency condensing coil 1. The coolant is pumped by pump 3 at high speed to form a local jet, actively disturbing and enhancing steam condensation; there is no need to start spray atomization, and pump 3 is maintained in normal operation; at the same time, the jet pump frequency is optimized according to historical coordinated condensation regulation data. Optimizing the pump frequency is equivalent to dynamically adjusting the pump speed to achieve energy efficiency optimization under the basic jet, taking into account both condensation efficiency and energy consumption control; when the condensation condition discrimination value is greater than or equal to the condensation threshold, it is determined that condensation is insufficient, and an enhanced condensation command is generated. The system activates the atomizing spray device on the high-efficiency condensing coil 1. This device, consisting of atomizing nozzles 7, electrically controlled valves 6, and an auxiliary pump, atomizes the coolant and sprays it into the steam chamber, enhancing heat exchange through jet condensation. It then enters a jet atomization-coordinated condensation mode. After entering enhanced condensation, the system continuously monitors the condensation condition judgment value. If this value falls below the condensation threshold, the system controls the electrically controlled valve 6 to close the atomizing nozzles 7, resuming jet condensation. The electrically controlled valve 6 precisely controls the start and stop of the nozzles 7, ensuring timely response to pressure and load changes and preventing excessive spraying that could lead to energy waste and liquid level runaway. The condensation condition judgment value is written into the coordinated condensation control database, recording the condensation mode switching and corresponding coordinated condensation control data throughout the process. Based on real-time feedback and historical response results, reinforcement learning and Bayesian optimization algorithms are used to optimize the steam chamber pressure benchmark, typical load, and condensation threshold. The reinforcement learning algorithm learns the optimal control strategy through interaction between the agent and the environment, while Bayesian optimization searches for the optimal parameters using a probabilistic model and data feedback, enabling continuous adaptive, refined, and intelligent condensation control.

[0046] like Figure 5The diagram shows the PID control condensation condition judgment process. It illustrates the main condition judgment and grading switching logic of two-phase flow jet atomization coordinated intelligent condensation. First, historical steam chamber pressure and server heat load baseline values ​​are extracted using the sliding time window quantile method, providing a dynamic and adaptive reference standard for subsequent judgments. Multiple types of sensors collect key physical parameters in real time, providing real-time data input for judging the condensation state and response actions. By fusing real-time pressure and load data, the current condensation condition judgment value is calculated, quantifying the matching degree between the current condensation capacity and load demand. The PID controller 9 compares the condensation condition judgment value with the dynamic condensation threshold to determine the current operating range. Based on the grading criteria: if the condensation condition judgment value is greater than or equal to the condensation threshold, it is determined that the load is too high and the condensation capacity is insufficient. The jet and atomized spray coordinated condensation mode is activated to improve condensation capacity and cope with high loads or extreme pressures. Otherwise, it is determined to be in a safe, energy-saving range, and only the jet condensation mode is activated to maintain basic operation, balancing energy efficiency and safety.

[0047] In this implementation scheme, dynamic matching and intelligent response of condensation capacity and load are achieved through graded switching based on condensation condition discrimination values, combined with the coordinated control of Venturi tube jet and atomizing spray equipment on high-efficiency condensing coil 1. Under normal operating conditions, an energy-saving jet condensation mode is adopted, and the pump frequency is optimized based on historical data to reduce energy consumption. Under high load and abnormal pressure conditions, the system switches to atomizing spray to enhance coordinated condensation, improving heat exchange efficiency and safety margin. Data from the entire process is written to the control database in real time, and reinforcement learning and Bayesian optimization algorithms are used to continuously optimize the pressure benchmark, typical load, and condensation threshold, ensuring that the criteria and parameters always closely match actual operating conditions.

[0048] Specifically, the process of entering the jet atomization-coordinated condensation mode and evaluating the intensity of the atomization spray based on the pre-processed coordinated condensation control data is as follows: Receive the enhanced condensation command and control the atomization spray equipment to perform atomization spray regulation; obtain the current steam chamber pressure and the steam chamber pressure reference value; divide the difference between the current steam chamber pressure and the steam chamber pressure reference value by the steam chamber pressure reference value to obtain the pressure over-limit ratio. The pressure over-limit ratio is used to quantify the proportion of the current pressure exceeding the normal operating reference and is an important criterion reflecting pressure risk and control requirements; add the pressure over-limit ratio to a constant and perform a natural logarithmic operation to obtain the pressure over-limit response value; obtain the current steam... The ratio of the current steam yield to the nozzle's rated maximum flow rate is used to determine the steam load capacity ratio. The steam yield reflects the ratio of the current server heat load to the latent heat of vaporization of the coolant, while the nozzle's rated maximum flow rate represents the maximum working capacity of the atomizing equipment. This ratio is used to determine the pressure level of the current load relative to the maximum atomization condensation capacity. Multiplying the pressure over-limit response value by the steam load capacity ratio yields the atomization spray coordinated control intensity value. This value comprehensively reflects the dual impact of pressure over-limit and heat load, serving as the core criterion for dynamically adjusting the atomization spray flow rate and its start / stop, and as the basis for subsequent graded control and safety protection decisions.

[0049] The specific formula for the intensity value of the synergistic regulation of atomized spray is as follows:

[0050] ;

[0051] In the formula, This represents the intensity value of coordinated control of atomized spray, which is used to dynamically adjust the intensity of atomized spray. It combines the current pressure exceeding the limit with the pressure of steam load on the spraying capacity, and quantifies it into an atomized spray control index that can directly drive the output of the controller. The larger the intensity value of coordinated control of atomized spray, the stronger the surge in condensation pressure and load, and the higher the spraying intensity is required. This indicates the current steam chamber pressure, reflecting the current load, condensation capacity, and safety margin. It represents the reference value of the steam chamber pressure, which serves as a reference baseline for pressure under normal operating conditions and dynamically adapts to different operating conditions. This indicates the current steam production rate and quantifies the actual heat dissipation requirements of the current heat load. This indicates the nozzle's rated maximum flow rate, a constant obtained from the nozzle's nameplate, reflecting the baseline of the physical limits of atomized spraying capabilities. It represents the pressure over-limit ratio, reflecting the degree to which the current pressure exceeds the reference pressure, and is the main physical basis for judging insufficient condensation and extreme load; This represents the pressure over-limit response value. A logarithmic function is used to respond to pressure over-limits, suppressing the effects of occasional anomalies. It is highly sensitive to continuous and significant pressure over-limits, enabling rapid enhancement of the spray response under overload conditions. It represents the steam load capacity ratio, which is the percentage of the current steam output relative to the maximum capacity of the atomizing spray, directly reflecting the load level and the spray limit distance.

[0052] This implementation scheme achieves intelligent and dynamic control of atomized spray intensity by comprehensively judging pressure over-limit response and steam load capacity. It not only sensitively reflects real-time changes in pressure and heat load, but also precisely adjusts spray flow and status according to equipment capacity, preventing energy waste and insufficient condensation. Overall, it improves safety, adaptability, and energy efficiency, achieving enhanced condensation and intelligent protection under high load and extreme operating conditions.

[0053] Specifically, based on the intensity assessment results of the atomized spray, the specific process of graded control of jet atomization synergistic condensation is as follows: the intensity value of atomized spray synergistic control is written into the synergistic condensation control database, and the intensity value of atomized spray synergistic control is compared with the multi-level intensity thresholds T1 and T2 in real time. The multi-level intensity thresholds T1 and T2 are the criteria boundaries for graded control, which can be dynamically optimized through historical operating conditions and self-learning algorithms to ensure that the switching of each state is sensitive and stable; when the intensity value of atomized spray synergistic control is less than T1, it enters the synergistic condensation energy-saving standby state; when the intensity value of atomized spray synergistic control is greater than or equal to T1 and less than T2, it enters the synergistic condensation normal adjustment state; when the intensity value of atomized spray synergistic control is greater than or equal to T2, it enters the synergistic condensation enhanced protection state; at the same time, corresponding condensation control measures are taken for different synergistic condensation states.

[0054] In this implementation plan, graded intelligent control is achieved by dynamically comparing the intensity value of the atomized spray with multi-level thresholds. Based on the actual load and operating conditions, it switches between various operating states—energy-saving standby, conventional adaptive adjustment, and enhanced protection—precisely matching the spray flow rate and pump speed. This graded response not only improves energy efficiency and condensation safety but also enhances risk prevention capabilities under extreme conditions, ensuring efficient, stable, and intelligent operation of the condenser under different load environments.

[0055] Specifically, the process of taking corresponding condensation control measures for different synergistic condensation states is as follows: When entering the synergistic condensation energy-saving standby state, maintain the minimum spray and jet flow rate and reduce the pump frequency. The minimum spray and jet flow rate is set by the minimum nozzle opening and the low-frequency operation of the pump, which can reduce energy consumption and prevent long-term stagnation from causing blockage of the pipes and nozzles 7; when the atomized spray synergistic control intensity value is less than T1 for a period of time exceeding the maximum allowable threshold, pump 3 operates intermittently, and nozzle 7 only performs periodic self-cleaning pulses. Intermittent operation means that pump 3 and nozzle 7 automatically start and stop according to the timer cycle, and self-cleaning pulses are used for short-term high-frequency operation. Frequent flushing prevents deposit blockage and ensures long-term reliable operation. When entering the synergistic condensation routine adjustment state, the atomized spray intensity is linearly mapped to the pump speed to increase the jet intensity. Linear mapping smoothly converts the intensity value into a pump speed control command according to a proportional relationship, achieving continuous, non-jumping adaptive adjustment. The spray flow rate is increased by controlling nozzle 7 via electronically controlled valve 6. Simultaneously, based on the server's heat load change rate (i.e., the magnitude of change in server heat load per unit time), real-time trend prediction allows for advance spray adjustment, avoiding condensation response lag during sudden load changes. This allows for early and gradual adjustment of the spray intensity and utilizes… The autoregressive moving average algorithm predicts future server heat load trends and optimizes the routine adjustment rhythm of coordinated condensation. This time-series data prediction model can predict load trends within the next 3 to 5 minutes based on historical load changes, assisting in spray and pump adjustment decisions. When entering the enhanced coordinated condensation state, nozzle 7 is fully open at its rated maximum flow rate and maximum jet flow rate. The full opening of nozzle 7 and maximum jet flow rate ensure that condensation capacity is at its highest guaranteed level under extreme high loads and abnormal pressures. Simultaneously, the backup drain valve is activated to prevent sudden increases in liquid level. This backup drain valve is a safety redundancy device and can be used in case of sudden increases in liquid level or failure of the main drain channel. Excess condensate is quickly drained to prevent overflow and equipment immersion. When the atomized spray coordinated control intensity value is greater than or equal to T2 for a period exceeding the maximum allowable threshold, an inspection work order is generated and an alarm is triggered to remind manual intervention for coolant overflow. The atomized spray coordinated control intensity value, nozzle flow rate, and the operating status of radiator 5 and condensation control equipment are recorded each time. The atomized spray coordinated control intensity value and multi-level intensity thresholds are continuously optimized using a Bayesian optimization algorithm. The Bayesian optimization algorithm can automatically analyze historical operating results and control effects, dynamically fine-tune the criteria and thresholds, and ensure that the control always maintains optimal adaptability and safety margin.

[0056] In this implementation plan, precise adaptive control of spray flow rate, jet intensity, and drainage is achieved by dynamically adopting graded regulation measures for different collaborative condensation states. This not only enables energy-saving operation and anti-clogging self-cleaning under low loads, and flexibly responds to load changes and optimizes adjustments in advance under normal conditions, but also ensures condensation and liquid level safety under extremely high loads, and improves operational efficiency through intelligent alarms and work order push notifications. Furthermore, relying on Bayesian optimization algorithms, various regulation parameters and thresholds can continuously learn and dynamically optimize.

[0057] Specifically, in the process of jet atomization-coordinated condensation control, the real-time evaluation of the comprehensive safety of liquid level and energy efficiency based on the pre-processed coordinated condensation control data is as follows: Real-time reception of the atomization spray coordinated control intensity value, nozzle flow rate, and the operating status of radiator 5 and condensation control equipment; during the jet atomization-coordinated condensation process, simultaneously acquiring the tank liquid level, the total energy consumption of the condensation control equipment, and the overall total operating energy consumption; based on a sliding time window, acquiring historical tank liquid level data and sorting it in ascending order; selecting the P5 quantile of the tank liquid level within the window as the lower limit of liquid level safety, and selecting the P95 quantile of the tank liquid level within the window as the upper limit of liquid level safety; using the P5 and P95 quantiles as the lower and upper limits of liquid level safety adaptable to different operating conditions, effectively shielding extreme abnormal data and making the liquid level judgment more robust; using the current tank liquid level and liquid level safety... The difference between the lower and lower limits, divided by the difference between the upper and lower safe limits of the liquid level, yields the relative safety margin value of the liquid level. This value reflects the current liquid level's position within the safe range; a value closer to 1 indicates the liquid level is near the upper limit, while a value closer to 0 indicates the liquid level is near the lower limit. This is an important reference for judging overflow and supply risks. The total energy consumption of the current condensation control equipment is divided by the total energy consumption of the overall operation to obtain the condensation energy consumption ratio. This ratio reflects the weight of the current condensation process in the total energy consumption; a high ratio indicates that the energy efficiency of the condensation process needs optimization. The condensation energy consumption ratio is multiplied by the energy efficiency weighting factor and added to the relative safety margin value of the liquid level to obtain the liquid level energy efficiency safety assessment value. The energy efficiency weighting factor can be dynamically set according to safety and energy consumption optimization needs. The liquid level energy efficiency safety assessment value integrates both liquid level and energy efficiency dimensions, achieving a comprehensive judgment of condensation safety and providing accurate decision-making basis for subsequent drainage and energy efficiency control.

[0058] The specific formula for the liquid level energy efficiency safety assessment value is as follows:

[0059] ;

[0060] In the formula, The liquid level energy efficiency safety assessment value is used to dynamically assess the safety of the condensate level and the energy efficiency level of the condensation process. It helps to determine whether the liquid level needs to be adjusted and energy consumption optimized, so as to achieve a comprehensive balance between safety and energy efficiency. The higher the liquid level energy efficiency safety assessment value, the closer the liquid level is to or beyond the upper limit, the higher the proportion of condensation energy consumption, and the higher the safety and energy efficiency risks. This indicates the current liquid level in the tank, reflecting the actual height of the condensate and directly determining whether there is a risk of liquid overflow. This indicates the safe lower limit of the liquid level, ensuring that the liquid level will not be too low and affect the normal cooling cycle. It is adaptively adjusted based on historical window data. This indicates the upper limit of the liquid level safety, representing the maximum allowable liquid level in the tank. It prevents overflow, malfunctions, and water ingress into the equipment. If the liquid level exceeds this value, the liquid must be drained immediately. This represents the total energy consumption of the current condensation control equipment. It is the sum of electrical energy consumed during the operation of equipment directly related to condensation enhancement and flow control, including pump 3, nozzle 7, electrically controlled valve 6, high-efficiency condensation coil 1, drain valve, and PID controller 9. Monitoring the actual electrical energy consumed during the condensation process is the core indicator for energy efficiency optimization. This represents the total energy consumption of the current overall operation, which is the sum of the power metering of all electrical equipment during operation. It includes not only the energy consumption of pumps 3, nozzles 7, electrically controlled valves 6, high-efficiency condenser coils 1, drain valves and PID controllers 9 in the condensation control link, but also the energy consumption of coolant tanks, radiators 5, temperature control modules, energy consumption acquisition modules, control units, sensor networks and all other operation-related electrical equipment, in order to reflect the total energy consumption. This represents the relative safety margin of the liquid level, quantifying the current position of the liquid level in the tank within the allowable range. The result is between [0,1], with a larger value indicating that it is closer to the upper limit and more dangerous. This indicates the proportion of condensation energy consumption, reflecting the current energy consumption of the condensation process to the total energy consumption. The larger the proportion, the heavier the burden on the condensation process and the lower the energy efficiency. The energy efficiency weighting factor represents the importance of balancing liquid level safety and energy efficiency constraints. It periodically counts the proportion of condensation energy consumption and the number of times the relative safety margin of liquid level exceeds the liquid level safety threshold within a historical window. With the dual objectives of minimizing energy consumption and optimizing liquid level safety, the energy efficiency weighting factor is fitted using a multi-objective genetic algorithm. When energy consumption is high but the liquid level is stable for a long period of time, the energy efficiency weighting factor is reduced. When the number of times the relative safety margin exceeds the liquid level safety threshold increases, the energy efficiency weighting factor is increased. The value range is between 0.1 and 0.5.

[0061] In this embodiment, Table 1 is a data table of liquid level energy efficiency safety assessment values. The lower limit of liquid level safety is set to 300, the upper limit of liquid level safety is set to 400, and the energy efficiency weighting factor is set to 0.25. The table records in detail the tank liquid level, total energy consumption of condensing control equipment, total energy consumption of overall operation, and liquid level energy efficiency safety assessment value at different times. Among them, the tank liquid level at time 1 is 340, the total energy consumption of condensing control equipment is 12.0, the total energy consumption of overall operation is 50.0, and the liquid level energy efficiency safety assessment value is 0.4600; the tank liquid level at time 2 is 310, the total energy consumption of condensing control equipment is 10.0, the total energy consumption of overall operation is 48.0, and the liquid level... The energy efficiency safety assessment value is 0.1521; the tank liquid level at time 3 is 380, the total energy consumption of the condensation control equipment is 16.0, the total energy consumption of the overall operation is 52.0, and the liquid level energy efficiency safety assessment value is 0.8769; the tank liquid level at time 4 is 395, the total energy consumption of the condensation control equipment is 18.0, the total energy consumption of the overall operation is 55.0, and the liquid level energy efficiency safety assessment value is 1.0318; the tank liquid level at time 5 is 350, the total energy consumption of the condensation control equipment is 14.0, the total energy consumption of the overall operation is 51.0, and the liquid level energy efficiency safety assessment value is 0.5686.

[0062] Table 1. Data Table of Liquid Level Energy Efficiency and Safety Assessment Values

[0063]

[0064] like Figure 6 The figure shows the trend of liquid level energy efficiency safety assessment values ​​and indicators. It illustrates the changing trends of the relative safety margin of liquid level, the weighted condensation energy consumption ratio, and the liquid level energy efficiency safety assessment values ​​at five different time points. The horizontal axis represents time, and the vertical axis represents each assessment value, labeled with different curves and symbols. (Based on Table 1 and...) Figure 6 It can be seen that when the liquid level in the tank is high, such as at times 3 and 4, both the liquid level energy efficiency safety assessment value and the relative safety margin of the liquid level rise rapidly, approaching or exceeding 1, entering the potential safety risk zone, requiring timely drainage and frequency reduction operations. Conversely, when the liquid level drops, the liquid level energy efficiency safety assessment value falls synchronously. The weighted condensation energy consumption ratio is consistently lower than the main line of the relative safety margin of the liquid level, serving as a correction factor for the liquid level energy efficiency safety assessment value and playing a fine-tuning role in the overall trend. This reflects the combined impact of the relative safety margin of the liquid level and the energy consumption ratio on the liquid level energy efficiency safety assessment value, and can be used for real-time monitoring and dynamic control of the condensation safety status.

[0065] In this implementation plan, the upper and lower limits of liquid level safety are adaptively set using the historical sliding window quantile method, and the relative margin of liquid level and the proportion of condensation energy consumption are integrated to dynamically calculate the liquid level energy efficiency safety assessment value. This not only accurately reflects the safety risk of the current liquid level and the energy efficiency level of the condensation process, but also effectively shields abnormal fluctuations, achieving coordinated intelligent judgment of liquid level and energy efficiency. It provides a scientific and real-time basis for safety decision-making, improving the safety, energy efficiency, and intelligent management capabilities of the condensation process.

[0066] Specifically, based on the comprehensive safety assessment results of liquid level and energy efficiency, the process of adjusting the drainage and condensation operation modes is as follows: The liquid level energy efficiency safety assessment value is compared with the energy efficiency threshold in real time. The energy efficiency threshold is the control limit for determining liquid level and energy efficiency safety, and can be adaptively optimized using historical operating data and algorithms to ensure neither overflow nor excessive energy consumption. When the liquid level energy efficiency safety assessment value is less than the energy efficiency safety threshold, it is determined that both the liquid level and energy consumption are in a safe state. The corresponding jet atomization synergistic condensation mode is maintained according to the atomization spray synergistic control intensity value, i.e., within the safe range, no excessive adjustment is required, thus improving operating efficiency and stability. Simultaneously, the liquid level energy efficiency safety assessment value is continuously monitored, and early warnings are issued. If the criterion shows a continuous upward trend, an early warning signal is issued in advance, reserving response time for subsequent control actions. When the liquid level energy efficiency safety assessment value is greater than or equal to the energy efficiency safety threshold, it is determined that there is a risk to the liquid level and condensation energy consumption, and the drainage valve is immediately activated to drain excess condensate. The condensate is drained back into the coolant tank 4. The drain valve is a key actuator in liquid cooling. It is opened by an electronic control signal to quickly reduce the tank level and prevent overflow and equipment damage. The nozzle opening and pump 3 speed are forcibly reduced to decrease the nozzle flow and jet intensity. By adjusting the operating parameters of the nozzle 7 and pump 3, the condensation load is rapidly downgraded to mitigate risks. If the time for which the liquid level energy efficiency safety assessment value is greater than or equal to the energy efficiency safety threshold exceeds the maximum allowable threshold, the draining force is increased and the atomization and jet power are downgraded. The maximum allowable threshold is the upper limit of tolerance for high-risk conditions. If the time limit is exceeded, the draining and energy consumption control are further enhanced to ensure safety boundaries. All energy efficiency safety assessment values ​​and control actions are written into the collaborative condensation control database, and the energy efficiency safety threshold is optimized and corrected based on historical feedback information. The historical feedback information includes actual overflow, alarms, and high energy consumption events. The energy efficiency threshold can be dynamically adjusted based on the feedback information to make the criteria adaptively optimized with actual operating conditions.

[0067] This implementation scheme achieves closed-loop management of drainage and condensation control through real-time discrimination and graded response of liquid level energy efficiency safety assessment values. It can maintain the optimal condensation mode within the safe range of liquid level and energy consumption, promptly initiate drainage and energy degradation measures when risks become critical, and continuously optimize judgment criteria and control strategies through data-driven threshold self-learning. This not only effectively prevents liquid level overflow and abnormal energy consumption but also significantly improves operational safety and energy efficiency management capabilities.

[0068] Reference Figure 2 As shown, the second aspect of the present invention provides an intelligent condensation control system for two-phase flow jet atomization synergy, applied to the aforementioned intelligent condensation control method for two-phase flow jet atomization synergy, comprising: a condensation data acquisition and preprocessing module for real-time acquisition of synergistic condensation control data and preprocessing the synergistic condensation control data; a condensation condition discrimination and condensation switching module for real-time discrimination of condensation conditions based on the preprocessed synergistic condensation control data and switching the jet atomization synergistic condensation mode according to the condensation conditions; an atomization spray enhanced condensation control module for entering the jet atomization synergistic condensation mode and evaluating the intensity of atomization spray based on the preprocessed synergistic condensation control data, and performing graded control of jet atomization synergistic condensation based on the intensity evaluation results; and a liquid level and energy efficiency safety closed-loop control module for real-time evaluation of liquid level and energy efficiency safety based on the preprocessed synergistic condensation control data during the jet atomization synergistic condensation control process, and adjusting the drainage and condensation operation modes based on the liquid level and energy efficiency safety evaluation results.

[0069] This implementation plan constructs a fully closed-loop, graded-response intelligent condensation control system through multi-module collaboration, including data acquisition and preprocessing, intelligent operating condition judgment, enhanced atomized spray control, and a closed-loop system for liquid level, energy efficiency, and safety. It can not only adaptively switch condensation modes and adjust spray intensity according to real-time operating conditions, but also dynamically evaluate and optimize the safety thresholds for liquid level and energy efficiency, achieving an organic unity of condensation capacity, energy efficiency management, and operational safety, thus improving energy efficiency, autonomy, and reliability in two-phase flow condensation scenarios.

[0070] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0071] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. As those skilled in the art will understand, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A smart condensation control method for synergistic atomization of two-phase flow jets, characterized in that, Includes the following steps: S1, collect collaborative condensation control data in real time, and perform data preprocessing on the collaborative condensation control data; S2, based on the pre-processed collaborative condensation control data, determines the condensation condition in real time and switches the jet atomization collaborative condensation mode according to the condensation condition; The specific process for real-time determination of condensation conditions based on preprocessed collaborative condensation control data is as follows: Based on the sliding time window, the historical steam chamber pressure is obtained, the historical steam chamber pressure data is sorted in ascending order, the P95 quantile is selected using the quantile method to obtain the steam chamber pressure benchmark value, and the current steam chamber pressure is divided by the steam chamber pressure benchmark value to obtain the steam chamber pressure margin ratio. Obtain the server heat load and the latent heat of vaporization constant of the coolant, and calculate the ratio of the current server heat load to the latent heat of vaporization constant of the coolant to obtain the current steam yield; Based on the sliding time window, historical server heat load is obtained, the historical server heat load data is sorted in ascending order, and the P95 quantile is selected using the quantile method to obtain the typical load. The ratio of the typical load to the latent heat of vaporization of the coolant is calculated to obtain the baseline steam yield. The current steam yield is divided by the baseline steam yield to obtain the active value of steam yield. The steam load response value is obtained by adding the active value of steam yield to a constant and performing a natural logarithmic operation; the condensation condition discrimination value is obtained by multiplying the steam chamber pressure margin ratio by the steam load response value. S3, enter the jet atomization synergistic condensation mode, and evaluate the intensity of atomization spray based on the pre-treated synergistic condensation control data. Based on the intensity evaluation results of atomization spray, perform graded control of jet atomization synergistic condensation. S4, during the jet atomization synergistic condensation control process, assesses the overall safety of liquid level and energy efficiency in real time based on the pre-treated synergistic condensation control data, and adjusts the drainage and condensation operation modes based on the overall safety assessment results of liquid level and energy efficiency.

2. The intelligent condensation control method for synergistic two-phase jet atomization according to claim 1, characterized in that, The specific process of real-time acquisition of coordinated condensation control data and data preprocessing of the coordinated condensation control data is as follows: By deploying multiple types of sensors at key nodes in the steam chamber, liquid cooling pipeline, and server (2) end, including steam chamber pressure sensor (8), liquid level sensor, nozzle flow sensor and intelligent power distribution unit, real-time collaborative condensation control data is collected. The collaborative condensation control data includes: steam chamber pressure, tank liquid level, server heat load, total energy consumption of condensation control equipment, total energy consumption of overall operation, nozzle flow, nozzle rated maximum flow and latent heat of vaporization of coolant. The collaborative condensation control data is denoised and smoothed using moving average and median filtering algorithms to remove sudden extreme values ​​and invalid data, and timestamp calibration and data alignment are standardized. For signals with missing or abnormal data, short-time historical interpolation, data source redundancy switching, and physical model verification methods are used for integrity correction. At the same time, the collaborative condensation control data is standardized and normalized, and consistency logic verification is performed. A collaborative condensation control database is established to store the collaborative condensation control data.

3. The intelligent condensation control method for synergistic two-phase jet atomization according to claim 1, characterized in that, The specific process of switching the jet atomization-coordinated condensation mode according to the condensation conditions is as follows: The condensation condition judgment value is calculated in real time. The PID controller (9) performs condensation regulation level switching according to the condensation condition judgment value: when the condensation condition judgment value is less than the condensation threshold, it is determined to be in normal condition and the jet condensation mode is activated: only the Venturi tube on the high-efficiency condensation coil (1) is driven to perform jet, without the need to start spray atomization, and the pump (3) is maintained in normal operation; at the same time, the jet pump frequency is optimized according to the historical collaborative condensation regulation data. When the condensation condition judgment value is greater than or equal to the condensation threshold, it is determined that the condensation is insufficient, and an enhanced condensation command is generated. The atomizing spray device on the high-efficiency condensing coil (1) is activated in conjunction with the jet atomization synergistic condensation mode. After entering the enhanced condensation mode, the condensation condition judgment value is continuously monitored. If the condensation condition judgment value falls back to below the condensation threshold, the electronic control valve (6) is controlled to close the atomizing nozzle (7) and jet condensation is restored. The condensation condition judgment value is written into the collaborative condensation control database. The condensation mode switching and corresponding collaborative condensation control data are recorded throughout the process. Based on real-time feedback and historical response results, reinforcement learning and Bayesian optimization algorithms are used to optimize the steam chamber pressure benchmark, typical load and condensation threshold.

4. The intelligent condensation control method for synergistic two-phase jet atomization according to claim 1, characterized in that, The specific process of entering the jet atomization-coordinated condensation mode and evaluating the intensity of the atomization spray based on the pre-processed coordinated condensation control data is as follows: The system receives a command to enhance condensation and controls the atomizing spray equipment to regulate atomizing spray. It acquires the current steam chamber pressure and the steam chamber pressure reference value, divides the difference between the two values ​​by the reference value to obtain the pressure over-limit ratio, adds this ratio to a constant, and performs a natural logarithmic operation to obtain the pressure over-limit response value. It also acquires the current steam yield and the nozzle's rated maximum flow rate, calculates the ratio of this ratio to obtain the steam load capacity ratio, and multiplies the pressure over-limit response value by the steam load capacity ratio to obtain the atomizing spray coordinated regulation intensity value.

5. The intelligent condensation control method for synergistic two-phase jet atomization according to claim 1, characterized in that, The specific process of graded control of jet atomization and synergistic condensation based on the intensity assessment results of atomized spray is as follows: The intensity value of the atomized spray coordinated control is written into the coordinated condensation control database, and the intensity value of the atomized spray coordinated control is compared with the multi-level intensity thresholds T1 and T2 in real time; when the intensity value of the atomized spray coordinated control is less than T1, it enters the coordinated condensation energy-saving standby state. When the intensity value of the coordinated control of atomized spray is greater than or equal to T1 and less than T2, it enters the normal adjustment state of coordinated condensation; when the intensity value of the coordinated control of atomized spray is greater than or equal to T2, it enters the enhanced protection state of coordinated condensation; at the same time, corresponding condensation control measures are taken for different coordinated condensation states.

6. The intelligent condensation control method for synergistic two-phase jet atomization according to claim 5, characterized in that, The specific process of taking corresponding condensation control measures for different synergistic condensation states is as follows: When entering the collaborative condensation energy-saving standby state, the minimum spray and jet flow rate is maintained and the pump frequency is reduced; when the time when the atomization spray collaborative control intensity value is less than T1 exceeds the maximum allowable threshold, the pump (3) adopts intermittent operation and the nozzle (7) only performs periodic self-cleaning pulses; When entering the synergistic condensation routine adjustment state, the jet intensity is increased by linearly mapping the atomized spray synergistic control intensity value to the pump speed, and the spray flow rate is increased by controlling the nozzle (7) through the electronic control valve (6); at the same time, the spray intensity is adjusted slowly in advance according to the server heat load change rate, and the future server heat load trend is predicted by the autoregressive moving average algorithm to optimize the synergistic condensation routine adjustment rhythm. When entering the collaborative condensation enhancement protection state, the nozzle (7) is fully opened at the rated maximum flow rate and the maximum jet flow rate is activated. At the same time, the standby drain valve is activated to prevent the liquid level from rising suddenly. When the time when the intensity value of the coordinated control of atomized spraying exceeds or equals T2 exceeds the maximum allowable threshold, an inspection work order is generated and an alarm push is initiated. Record the intensity value of atomization spray coordination control, nozzle flow rate and the operating status of radiator (5) and condensation control equipment for each atomization spray coordination control, and continuously optimize the intensity value of atomization spray coordination control and multi-level intensity thresholds through Bayesian optimization algorithm.

7. The intelligent condensation control method for synergistic atomization of two-phase flow jets according to claim 1, characterized in that, The specific process of real-time evaluation of liquid level and overall energy efficiency safety based on pre-treated synergistic condensation control data during jet atomization is as follows: Real-time reception of atomization spray coordinated control intensity value, nozzle flow rate and radiator (5) and condensation control equipment operation status. During the jet atomization coordinated condensation process, the tank liquid level, total energy consumption of condensation control equipment and total overall operating energy consumption are obtained simultaneously. Based on the sliding time window, historical tank liquid level data is obtained and sorted in ascending order. The P5 quantile of the tank liquid level in the window is selected as the lower limit of liquid level safety, and the P95 quantile of the tank liquid level in the window is selected as the upper limit of liquid level safety. The difference between the current tank liquid level and the lower limit of liquid level safety is divided by the difference between the upper limit of liquid level safety and the lower limit of liquid level safety to obtain the relative safety margin value of liquid level. The condensation energy consumption ratio is obtained by dividing the current total energy consumption of the condensation control equipment by the current total operating energy consumption. The condensation energy consumption ratio is then multiplied by the energy efficiency weighting factor and added to the liquid level relative safety margin value to obtain the liquid level energy efficiency safety assessment value.

8. The intelligent condensation control method for synergistic atomization of two-phase flow jets according to claim 1, characterized in that, The specific process of adjusting the drainage and condensation operation modes based on the comprehensive safety assessment results of liquid level and energy efficiency is as follows: The system compares the liquid level energy efficiency safety assessment value with the energy efficiency threshold in real time. When the liquid level energy efficiency safety assessment value is less than the energy efficiency safety threshold, it is determined that the liquid level and energy consumption are both in a safe state. The corresponding jet atomization condensation mode is maintained according to the intensity value of the atomization spray synergistic control. At the same time, the liquid level energy efficiency safety assessment value is continuously monitored to provide early warning. When the liquid level energy efficiency safety assessment value is greater than or equal to the energy efficiency safety threshold, it is determined that there is a risk to the liquid level and condensation energy consumption. The drain valve is immediately activated to drain the excess condensate back to the coolant tank (4), and the nozzle opening and pump speed are forcibly reduced to reduce the nozzle flow rate and jet intensity. If the liquid level energy efficiency safety assessment value is greater than or equal to the energy efficiency safety threshold for more than the maximum allowable threshold, the draining force is increased and the atomization and jet power are downgraded. All energy efficiency safety assessment values ​​and control actions are written into the collaborative condensation control database, and the energy efficiency safety thresholds are optimized and corrected based on historical feedback information.

9. A two-phase jet atomization-coordinated intelligent condensation control system, employing the two-phase jet atomization-coordinated intelligent condensation control method as described in any one of claims 1-8, characterized in that, include: The condensation data acquisition and preprocessing module is used to acquire collaborative condensation control data in real time and perform data preprocessing on the collaborative condensation control data. The condensation condition judgment and condensation switching module is used to judge the condensation condition in real time based on the pre-processed collaborative condensation control data, and switch the jet atomization collaborative condensation mode according to the condensation condition. The atomization spray enhanced condensation control module is used to enter the jet atomization synergistic condensation mode, and evaluate the intensity of atomization spray based on the pre-processed synergistic condensation control data. Based on the intensity evaluation results of atomization spray, the jet atomization synergistic condensation is controlled in stages. The liquid level and energy efficiency safety closed-loop control module is used to evaluate the overall safety of liquid level and energy efficiency in real time based on the pre-treated collaborative condensation control data during the jet atomization collaborative condensation control process. At the same time, based on the overall safety assessment results of liquid level and energy efficiency, it adjusts the drainage and condensation operation modes.

Images