Condenser dynamic heat exchange regulation and control method and system based on multi-dimensional perception and AI

Abstract

The invention discloses a condenser dynamic heat exchange regulation and control method and system based on multi-dimensional perception and AI, and relates to the technical field of intelligent heat exchange regulation and control. The condenser dynamic heat exchange regulation and control method and system based on multi-dimensional perception and AI comprises the steps that S1, condenser heat exchange perception data are collected and preprocessed; s2, performing two rounds of data screening on the sampling points, realizing emergency regulation and control instruction issuing and abnormal point data processing, and constructing an effective condenser operation data set; s3, the heat exchange adaptation degree of the condenser is evaluated, the load state category is recognized, and a regulation and control instruction is output; and S4, the regulation and control instruction is executed, execution feedback data is collected, the heat exchange execution performance of the condenser is evaluated, whether the current regulation and control instruction needs to be adjusted or not is judged based on the evaluation result, and the execution feedback data and the evaluation result are returned for threshold value updating and regulation and control strategy optimization. The problems that due to fixed heat exchange control of the condenser, phase change is limited under load change, and energy consumption is too high are solved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent heat exchange control technology, and specifically to a condenser dynamic heat exchange control method and system based on multi-dimensional perception and AI. Background Art

[0002] With the widespread adoption of high-performance computing, high-density packaging, and intelligent terminal devices, the heat dissipation capacity of electronic systems has become a key factor restricting their stable operation and energy efficiency optimization. The condenser, as the core heat dissipation unit, has a significant impact on chip temperature control and overall system power consumption due to its operating status. Traditional condenser control strategies are primarily based on fixed thresholds and rule-driven approaches, lacking the ability to adapt to changes in the operating environment in real time. This can easily lead to problems such as insufficient heat dissipation, excessive energy consumption, or resource waste, making it difficult to meet the refined thermal management requirements under current complex operating conditions.

[0003] For example, the invention with publication number CN115293373A provides a method, device, and electronic device for determining a condenser maintenance strategy. The method includes: obtaining the exhaust data and condensate data of the steam turbine, fan data, and the heat exchange area of ​​the condenser; the exhaust data includes the temperature after the heat exchange of the last stage exhaust steam and the initial temperature of the last stage exhaust steam; based on the exhaust data and condensate data of the steam turbine, determining the steam heat release of the steam turbine; based on the steam heat release, fan data, the temperature after the heat exchange of the last stage exhaust steam, and the initial temperature of the last stage exhaust steam, determining the heat exchange temperature difference of the condenser; based on the steam heat release, heat exchange temperature difference, and heat exchange area, determining the overall heat exchange coefficient; and based on the overall heat exchange coefficient, determining the maintenance strategy of the condenser. The present invention can determine a reasonable condenser maintenance strategy based on the operating conditions of the air-cooled thermal power unit to ensure the normal operation of the air-cooled thermal power unit.

[0004] For example, the invention with announcement number CN104166794B discloses a nuclear power plant condenser characteristic test acquisition and analysis system. This acquisition and analysis system includes: a data acquisition system for acquiring the data required for the condenser characteristic test; an IMP acquisition system for collecting and processing the acquired seawater inlet and outlet temperatures, condensate temperature, and back pressure data; and a calculation and analysis system for reading data from the IMP acquisition system in real time, receiving user-input data on nuclear island thermal power, generator power, and circulating water salinity, and calculating and displaying the test results based on the condenser design parameters and the received data. The implementation of the technical solution of the present invention is simple to operate and provides accurate analysis, ensuring the analysis and evaluation of the condenser's thermal performance. This ensures the condenser's good operating condition through maintenance and improvement, thereby improving the unit's safe and economical operation.

[0005] However, although the above technical solution can improve the safety and reliability of condenser operation in specific scenarios, it still has the following shortcomings: lack of real-time perception, fusion and dynamic processing capabilities of operating data; lack of dynamic control mechanism based on AI strategy optimization, making it difficult to achieve heat dissipation efficiency guarantee under high load and energy consumption compression under low load; the control logic of condenser heat exchange adaptability evaluation, load identification and closed-loop linkage of control instructions has not been established, and the system intelligence and adaptability are insufficient.

[0006] Therefore, in response to the above problems, there is an urgent need for a dynamic heat exchange control method and system for condensers based on multi-dimensional perception and AI. Summary of the Invention

[0007] Technical problems solved In response to the shortcomings of the existing technology, the present invention provides a dynamic heat exchange control method and system for a condenser based on multi-dimensional perception and AI, which solves the problem of limited phase change and excessive energy consumption under load changes caused by fixed heat exchange control of the condenser.

[0008] Technical Solution To achieve the above objectives, the present invention is implemented through the following technical solutions: a dynamic heat exchange control method and system for a condenser based on multi-dimensional perception and AI, including: S1, collecting condenser heat exchange perception data, and performing time synchronization, noise separation, physical consistency verification and normalized mapping processing on the condenser heat exchange perception data; S2, based on the preprocessed condenser heat exchange perception data, performing two rounds of data screening on the sampling points, realizing the issuance of emergency control instructions and abnormal point data processing, and constructing an effective condenser operation data set; S3, evaluating the condenser heat exchange adaptation degree based on the effective condenser operation data set, and extracting the chip temperature deviation ratio and the box air pressure deviation ratio to identify the load state category; inputting the heat exchange adaptation evaluation results, load state category and effective condenser operation data into the reinforcement learning decision model, and outputting the control instruction; S4, executing the control instruction and collecting execution feedback data, evaluating the condenser heat exchange execution performance, judging whether the current control instruction needs to be adjusted based on the evaluation results, and transmitting the execution feedback data and the evaluation results back for threshold update and control strategy optimization.

[0009] Furthermore, the specific steps of collecting the condenser heat exchange perception data and performing time series synchronization, noise separation, physical consistency verification and normalized mapping processing on the condenser heat exchange perception data are as follows: collecting the condenser heat exchange perception data, the condenser heat exchange perception data includes chip temperature, box air pressure, cooling coil inlet water temperature, cooling coil outlet water temperature, cooling water density, cooling water flow, cooling water specific heat capacity, pump power, box saturated steam pressure; through multi-channel timestamp comparison and sampling delay compensation method, the condenser heat exchange perception data is synchronized across sensors and the sampling rate is unified; by combining Kalman filtering The multi-scale signal reconstruction method based on wavelet decomposition is used to eliminate measurement noise and separate periodic disturbances of the condenser heat exchange perception data; a heat flow integrated consistency verification mechanism is established based on the energy conservation equation and the pressure difference flow coupling model to identify and eliminate data that does not conform to physical laws in the condenser heat exchange perception data; through multi-dimensional feature encoding and transmission protocol adaptive encapsulation methods, the data format of the condenser heat exchange perception data is standardized and the interface compatibility conversion is processed; through physical quantity normalization mapping and dynamic interval stretching algorithm, the unit conversion and feature scale normalization of the condenser heat exchange perception data are carried out.

[0010] Furthermore, based on the pre-processed condenser heat exchange perception data, two rounds of data screening are performed on the sampling points to realize the issuance of emergency control instructions and abnormal point data processing. The specific steps for constructing an effective condenser operation data set are as follows: extract the pre-processed condenser heat exchange perception data, and perform the first round of data screening on each sampling point: when the chip temperature exceeds the chip temperature upper limit threshold, the box air pressure exceeds the air pressure upper limit threshold, or the cooling water flow is lower than the flow lower limit threshold, the data is judged to be abnormal and emergency control is triggered, and the target cooling coil inlet water temperature and the target pump power are adjusted to the emergency set value and issued to the actuator, and the remaining sampling points enter the second screening; for the sampling points entering the second screening, the cooling coil inlet and outlet water temperature difference is calculated according to the difference between the cooling coil inlet water temperature and the cooling coil outlet water temperature, and the cooling coil inlet and outlet water temperature difference is calculated in the sliding The mean and standard deviation of chip temperature, box air pressure, cooling water flow and cooling coil inlet and outlet water temperature difference are calculated within the dynamic window to evaluate the degree of condenser operating condition abnormality and obtain the condenser operating condition abnormality evaluation value; the condenser operating condition abnormality evaluation value and the operating condition abnormality threshold are compared in real time. When the condenser operating condition abnormality evaluation value exceeds the abnormality score threshold, the corresponding sampling point is determined to be an abnormal point; if it is an isolated abnormal point, it is replaced by the sliding average of the condenser heat exchange perception data of the adjacent sampling points; if three or more consecutive sampling points are determined to be abnormal points and the corresponding parameter collection source is the same sensor, the backup sensor collection value is switched; otherwise, the corresponding sampling point is determined to be a normal point and the original data is retained; the condenser heat exchange perception data retained after two rounds of data screening are used to construct an effective condenser operation data set in the order of sampling points.

[0011] Furthermore, the specific steps for evaluating the abnormality of the condenser working condition and obtaining the abnormality evaluation value of the condenser working condition are as follows: divide the square of the difference between the chip temperature and the chip temperature mean by the square of the chip temperature standard deviation plus one to obtain the temperature deviation term; divide the square of the difference between the box air pressure and the box air pressure mean by the square of the box air pressure standard deviation plus one to obtain the air pressure deviation term; divide the square of the difference between the cooling water flow rate and the cooling water flow mean by the square of the cooling water flow standard deviation plus one to obtain the flow deviation term; divide the square of the difference between the inlet and outlet water temperature of the cooling coil and its mean by the square of the standard deviation of the cooling water flow rate plus one to obtain the flow deviation term. Add one to the square of the temperature difference standard deviation to obtain the temperature difference deviation term; add the temperature deviation term, air pressure deviation term, flow deviation term and temperature difference deviation term and take the square root to obtain the multi-parameter normalized deviation sum; multiply the difference between the chip temperature and the target chip safety temperature by the difference between the box air pressure and the box saturated steam pressure, take the absolute value and divide it by the product of the target chip safety temperature and the box saturated steam pressure to obtain the temperature-pressure interaction correction term; add one to the temperature-pressure interaction correction term to obtain the correction factor; multiply the multi-parameter normalized deviation sum by the correction factor to obtain the condenser operating condition abnormality assessment value.

[0012] Furthermore, the heat transfer adaptability of the condenser is evaluated based on the effective condenser operation data set, and the chip temperature deviation ratio and the box air pressure deviation ratio are extracted. The specific steps for identifying the load state category are as follows: Extract the effective condenser operation data set to evaluate the heat transfer adaptability of the condenser: multiply the cooling water density, cooling water flow rate, and cooling water specific heat capacity by the cooling coil inlet and outlet water temperature difference to obtain the heat transfer power component; take the larger value of the difference between the box saturated steam pressure and the box air pressure and the minimum term to obtain the pressure margin component; divide the heat transfer power component by the pressure margin component to obtain the pressure-corrected heat transfer power; The difference between the chip temperature and the target chip safety temperature is divided by the target chip safety temperature to obtain the relative temperature deviation coefficient; the pressure-corrected heat exchange power is multiplied by the relative temperature deviation coefficient to obtain the condenser heat exchange adaptation evaluation value; the chip temperature mean and the box air pressure mean are extracted and compared with the chip temperature threshold and the box air pressure threshold respectively to obtain the chip temperature deviation ratio and the box air pressure deviation ratio; the condenser heat exchange adaptation evaluation value, the chip temperature deviation ratio and the box air pressure deviation ratio are composed into a feature vector, which is input into the BP neural network classifier to output the current load state category, including high load, medium load and low load.

[0013] Furthermore, the heat exchange adaptation evaluation results, load state category and effective condenser operation data are input into the reinforcement learning decision model, and the specific steps for outputting the control instructions are as follows: based on the historical effective condenser operation data, the Q-Learning algorithm is used to construct the reinforcement learning decision model, the cold target cooling coil inlet water temperature adjustment amount and the target pump power adjustment amount are set as the action set, the load state category, condenser heat exchange adaptation evaluation value and effective condenser operation data are set as the state set, and the reinforcement learning decision model training is completed with maximizing the heat dissipation efficiency under high load conditions and minimizing the energy consumption under low load conditions as the reward function; the current load state category, condenser heat exchange adaptation evaluation value and effective condenser operation data are input into the reinforcement learning decision model as decision input parameters, and the control instructions are output: reduce the cold target cooling coil inlet water temperature and increase the target pump power under high load conditions; increase the cold target cooling coil inlet water temperature and reduce the target pump power under low load conditions; maintain the current cold target cooling coil inlet water temperature and target pump power under medium load conditions.

[0014] Furthermore, the specific steps of executing the control instructions and collecting execution feedback data are as follows: receiving the control instructions, sending the target cooling coil inlet water temperature adjustment amount to the cooling water temperature control execution unit, sending the target pump power adjustment amount to the pump speed control unit, and driving the control execution for real-time adjustment; during the control execution process, collecting the actual value of the cooling coil inlet water temperature after execution, the actual value of the cooling coil outlet water temperature after execution, the actual value of the pump power after execution, the actual execution response time and the actual total energy consumption during the control cycle, and constructing an execution feedback data set.

[0015] Furthermore, the specific steps for evaluating the heat exchange execution performance of the condenser are as follows: based on the execution feedback data set, the cooling water density, cooling water flow rate and cooling water specific heat capacity are multiplied together, and then multiplied by the difference between the actual value of the cooling coil inlet water temperature after execution and the actual value of the cooling coil outlet water temperature after execution to obtain the heat exchange power component; the actual execution response time is divided by the benchmark response time and added one to obtain the response time correction term; the absolute value of the difference between the actual value of the cooling coil inlet water temperature after execution and the target cooling coil inlet water temperature is calculated, and the absolute value of the difference between the actual value of the pump power after execution and the target pump power is calculated, the two absolute values ​​are added together and then added one to obtain the execution deviation correction term; the actual total energy consumption, the response time correction term and the execution deviation correction term in the control period are multiplied together to obtain the total corrected energy consumption; the heat exchange power component is divided by the total corrected energy consumption to obtain the heat exchange execution performance evaluation value.

[0016] Furthermore, based on the evaluation results, it is determined whether the current control instruction needs to be adjusted, and the execution feedback data and the evaluation results are transmitted back. The specific steps for threshold update and control strategy optimization are as follows: the heat exchange execution performance evaluation value is compared with the execution performance threshold in real time. When the heat exchange execution performance evaluation value is greater than or equal to the execution performance threshold, the current control strategy is maintained for the next control cycle. When the heat exchange execution performance evaluation value is less than the execution performance threshold, the cooling coil inlet water temperature deviation and pump power deviation are determined based on the correspondence between the actual value of the cooling coil inlet water temperature after execution and the target cooling coil inlet water temperature, and the actual value of the pump power after execution and the target pump power recorded in the execution feedback data set. According to the direction and magnitude of the deviation, a corrected cold target cooling coil inlet water temperature and target pump power instruction are generated and issued to the actuator. Regardless of whether the strategy is corrected, the execution feedback data set and the heat exchange execution performance evaluation value are transmitted back to the data screening link and the intelligent decision-making link respectively. The temperature threshold, air pressure threshold, and flow threshold are updated in the data screening link. The cold target cooling coil inlet water temperature adjustment strategy and the target pump power adjustment strategy are optimized in the intelligent decision-making link to achieve closed-loop operation of dynamic heat exchange control.

[0017] The second aspect of the present invention provides a condenser dynamic heat exchange control system based on multi-dimensional perception and AI, including: a data perception preprocessing module, an abnormal screening emergency control module, an intelligent state judgment decision generation module and an execution feedback closed-loop optimization module, wherein: the data perception preprocessing module is used to collect condenser heat exchange perception data, and perform time series synchronization, noise separation, physical consistency verification and normalization mapping processing on the condenser heat exchange perception data; the abnormal screening emergency control module is used to perform two rounds of data screening on the sampling points based on the preprocessed condenser heat exchange perception data, realize the issuance of emergency control instructions and abnormal point data processing, and build an effective condenser. Condenser operation data set; intelligent state judgment decision generation module, used to evaluate the condenser heat exchange adaptation degree based on the effective condenser operation data set, and extract the chip temperature deviation ratio and the box air pressure deviation ratio to identify the load state category; the heat exchange adaptation evaluation results, load state category and effective condenser operation data are input into the reinforcement learning decision model, and the control instructions are output; the execution feedback closed-loop optimization module is used to execute the control instructions and collect execution feedback data, evaluate the condenser heat exchange execution performance, determine whether the current control instructions need to be adjusted based on the evaluation results, and send back the execution feedback data and evaluation results for threshold update and control strategy optimization.

[0018] Beneficial effects The present invention has the following beneficial effects: (1) This condenser dynamic heat exchange control method and system based on multi-dimensional perception and AI realizes real-time identification and emergency control response of abnormal condenser operating conditions by setting safety thresholds of chip temperature, box air pressure and cooling water flow, combined with a two-round data screening mechanism. It can automatically issue control instructions when the sampling data exceeds the limit, effectively preventing problems such as heat exchange failure, abnormal energy consumption or system overheating, and improving the safety and stability of the system.

[0019] (2) This condenser dynamic heat exchange control method and system based on multi-dimensional perception and AI, by constructing a condenser heat exchange adaptation evaluation model and combining the deviation ratio of chip temperature and box air pressure, uses a BP neural network to achieve high-precision intelligent identification of the current load state, opens up the feedback path between condenser operation and calculated load, provides targeted input for subsequent heat exchange strategies, and realizes the transition from static control to state perception drive.

[0020] (3) The dynamic heat exchange control method and system for condensers based on multi-dimensional perception and AI uses the Q-Learning reinforcement learning algorithm to build a decision model. With the optimization goals of maximizing heat dissipation efficiency under high load conditions and minimizing energy consumption under low load conditions, it trains and outputs the optimal cooling coil water temperature and pump power control instructions, realizes self-learning and dynamic evolution of the strategy, and significantly improves the energy efficiency ratio and intelligent control level of the condenser.

[0021] (4) The condenser dynamic heat exchange control method and system based on multi-dimensional perception and AI builds a heat exchange execution performance evaluation model to quantitatively judge the control execution effect in real time, and transmits the execution feedback data back to the data screening and decision-making link to complete the dynamic update of thresholds such as temperature, air pressure and flow and the optimization iteration of the strategy model, forming a closed-loop adaptive heat exchange control mechanism to ensure the long-term stable and efficient operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a flow chart of the condenser dynamic heat exchange control method based on multi-dimensional perception and AI; Figure 2 This is the structural diagram of the condenser dynamic heat exchange control system based on multi-dimensional perception and AI; Figure 3 It is the trend chart of the condenser heat exchange adaptation evaluation value; Figure 4 This is the schematic diagram of the condenser dynamic heat exchange control system based on multi-dimensional perception and AI.

[0023] In the figure, 1. Saturated steam pressure sensor; 2. Cooling coil; 3. Condenser; 4. Server; 5. Temperature sensor; 6. Pump; 7. Coolant reservoir. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] See also Figures 1-4 , an embodiment of the present invention provides a technical solution: a dynamic heat exchange control method and system for a condenser based on multi-dimensional perception and AI, including: S1, collecting condenser heat exchange perception data, and performing time synchronization, noise separation, physical consistency verification and normalization mapping processing on the condenser heat exchange perception data; S2, based on the preprocessed condenser heat exchange perception data, performing two rounds of data screening on the sampling points, realizing the issuance of emergency control instructions and abnormal point data processing, and constructing an effective condenser operation data set; S3, evaluating the heat exchange adaptation degree of condenser 3 based on the effective condenser operation data set, and extracting the chip temperature deviation ratio and the box air pressure deviation ratio to identify the load state category; inputting the heat exchange adaptation evaluation result, the load state category and the effective condenser operation data into the reinforcement learning decision model, and outputting the control instruction; S4, executing the control instruction and collecting execution feedback data, evaluating the heat exchange execution performance of condenser 3, judging whether the current control instruction needs to be adjusted based on the evaluation result, and transmitting the execution feedback data and the evaluation result back for threshold update and control strategy optimization.

[0026] Specifically, the specific steps of collecting condenser heat exchange perception data and performing time series synchronization, noise separation, physical consistency verification and normalized mapping processing on the condenser heat exchange perception data are as follows: collecting condenser heat exchange perception data, the condenser heat exchange perception data includes chip temperature, box air pressure, cooling coil inlet water temperature, cooling coil outlet water temperature, cooling water density, cooling water flow, cooling water specific heat capacity, pump power, box saturated steam pressure, each perception data comes from an independent sensor channel and has a corresponding timestamp and sampling frequency label; through multi-channel timestamp comparison and sampling delay compensation method, combined with the sampling start time and sampling period of each data channel, cross-sensor time series synchronization and sampling rate unified processing are performed on each condenser heat exchange perception data to ensure that various perception data are coordinated and aligned under a unified time base; through a multi-scale signal reconstruction method combining Kalman filtering and wavelet decomposition, multi-dimensional filtering and disturbance separation processing are performed on each condenser heat exchange perception data respectively to effectively suppress high-frequency measurement noise and extract stable heat exchange trends; based on energy conservation A heat flow integration consistency verification mechanism is established by coupling the constant equation with the pressure difference flow model. The cooling water density data, cooling water flow data and cooling water specific heat capacity data are multiplied together, and then the theoretical heat exchange is estimated based on the temperature difference between the cooling coil inlet water temperature data and the cooling coil outlet water temperature data. The theoretical heat exchange is then compared and analyzed with the load heat flow derived from the chip temperature data and the box air pressure data. Abnormal sampling points inconsistent with the thermal balance are identified, and operations such as chip temperature data removal, box air pressure data removal, and cooling water flow data removal are performed to ensure that the input data follows physical laws. Through multi-dimensional feature encoding and transmission protocol adaptive packaging methods, each condenser heat exchange perception data is encoded in a unified format and a data packaging structure that can be adapted to different communication interfaces is constructed to achieve lossless intercommunication and parallel processing of cross-system data. Through physical quantity normalization mapping and dynamic interval stretching algorithm, unit conversion and feature scale normalization are performed on each condenser heat exchange perception data, and a standardized perception vector group is output as an effective input for subsequent heat exchange control algorithm and state judgment model.

[0027] In this implementation scheme, by sequentially performing cross-sensor timing synchronization, multi-scale noise separation, physical consistency verification, format standardization packaging and feature normalization mapping processing on chip temperature data, box air pressure data, cooling coil inlet water temperature data, cooling coil outlet water temperature data, cooling water density data, cooling water flow data, cooling water specific heat capacity data, pump power data and box saturated steam pressure data, data quality control can be achieved in terms of time alignment, numerical accuracy, physical rationality, interface matching and dimensional unification, providing a high-confidence data basis for subsequent abnormality screening, state identification, control generation and feedback evaluation, thereby improving the accuracy, stability and adjustability of the heat exchange control process.

[0028] Specifically, based on the preprocessed condenser heat exchange sensing data, two rounds of data screening are performed for each sampling point to implement emergency control command issuance and abnormal point data processing. The specific steps for constructing an effective condenser operation data set are as follows: extract the condenser heat exchange sensing data after cross-sensor time series synchronization processing, multi-scale noise separation processing, physical consistency verification processing, data format standardization processing, and physical quantity normalization mapping processing. For each sampling point, chip temperature data, box air pressure data, and cooling water flow data are extracted in sequence to perform the first round of data screening operations. When the chip temperature exceeds the chip temperature upper threshold, the box air pressure exceeds the box air pressure upper threshold, and the cooling water flow is lower than the cooling water flow lower threshold, the data corresponding to the sampling point is determined to be abnormal, and the emergency control process is immediately triggered. The cold target cooling coil inlet water temperature is set to the emergency safety value corresponding to the cooling water temperature, and the target pump power is set to the emergency output value within the pump power control range. These two emergency setting values ​​are packaged in instructions and issued to the corresponding actuator. The remaining sampling points that have not triggered abnormalities enter the second screening process. The actuator is a key component for implementing control strategies, receiving control instructions and completing corresponding physical adjustments. The actuator includes a cooling water temperature control actuator and a pump speed control unit. The cooling water temperature control actuator precisely adjusts the cooling medium temperature based on the target cooling coil inlet water temperature adjustment; the pump speed control unit controls the pump speed based on the target pump power adjustment, thereby adjusting the cooling water circulation rate. The actuator's response speed and adjustment accuracy directly impact the actual implementation of the heat exchange control strategy and are the core execution components for achieving dynamic heat exchange optimization.For the sampling points entering the second screening process, the cooling coil inlet and outlet water temperature difference corresponding to the current sampling point is calculated based on the real-time difference between the cooling coil inlet water temperature and the cooling coil outlet water temperature; based on the preset sliding window, the sliding mean and sliding standard deviation of the chip temperature, box air pressure, cooling water flow and the cooling coil inlet and outlet water temperature difference are calculated respectively, and the statistical feature space of the thermal parameters of condenser 3 is constructed. Combined with the multi-parameter normalized deviation and the temperature-pressure interaction correction factor, the abnormal degree of the working condition of condenser 3 at the sampling point is evaluated, and finally the condenser working condition abnormality evaluation value corresponding to the sampling point is obtained; the condenser working condition abnormality evaluation value is compared with the working condition abnormality scoring threshold in real time. When the condenser working condition abnormality evaluation value exceeds the currently set working condition abnormality scoring threshold, the sampling point is determined to be an abnormal point; if the abnormal point shows an isolated distribution feature in the time series, the chip of the adjacent sampling points before and after is used. The abnormal point is replaced by the sliding average of four types of condenser heat exchange sensing data: temperature, box air pressure, cooling water flow and cooling coil inlet and outlet water temperature difference; if three or more consecutive sampling points are judged as abnormal points, and their corresponding abnormal parameter data all come from the same sensor channel, it is judged that the acquisition source is unstable, and the corresponding backup sensor acquisition value is immediately switched to, and the sampling sequence number is retained; otherwise, the sampling point is judged to be a valid sampling point, and its original condenser heat exchange sensing data is retained without correction; finally, the nine types of condenser heat exchange sensing data retained after two rounds of screening, including chip temperature, box air pressure, cooling water flow, cooling coil inlet water temperature, cooling coil outlet water temperature, cooling water density, cooling water specific heat capacity, pump power and box saturated steam pressure, are arranged in sequence according to sampling time to construct an effective condenser operation data set for subsequent heat exchange adaptation evaluation and control instruction generation.

[0029] In this implementation, two rounds of point-by-point screening of the condenser heat exchange sensing data are performed. An emergency control trigger mechanism is constructed by combining the chip temperature upper threshold, the cabinet air pressure upper threshold, and the cooling water flow lower threshold. A sliding mean and sliding standard deviation calculation method is introduced to form an assessment value for condenser operating abnormalities, effectively identifying sudden abnormal operating conditions and data source stability issues. This method can accurately replace isolated abnormal sampling point data and dynamically switch the instability sensor acquisition channel, ultimately screening out a valid condenser operating data set that meets the statistical laws of Condenser 3's thermal characteristics. This lays an accurate and reliable data foundation for subsequent heat exchange adaptation evaluation and control instruction generation, significantly improving the interference resistance of data processing and the execution reliability of Condenser 3 control.

[0030] Specifically, the specific steps for evaluating the abnormality of the working condition of condenser 3 and obtaining the abnormal working condition evaluation value of condenser 3 are as follows: square the difference between the chip temperature corresponding to the current sampling point and the mean of all chip temperature sample values ​​in the sliding window, and divide the square difference by the sum of the square of the standard deviation of the chip temperature sample value and the constant term one to obtain the temperature deviation term; square the difference between the box air pressure corresponding to the sampling point and the mean of the box air pressure samples in the sliding window, and divide the square difference by the sum of the square of the standard deviation of the box air pressure samples and the constant term one to obtain the air pressure deviation term; The square of the difference between the cooling water flow rate corresponding to the sampling point and the mean cooling water flow rate in the sliding window is taken, and the square difference is divided by the sum of the square of the standard deviation of the cooling water flow rate samples and the constant term 1 to obtain the flow deviation term; the difference between the cooling coil inlet water temperature and the cooling coil outlet water temperature is obtained to obtain the current cooling coil inlet and outlet water temperature difference, and the square of the difference with the mean of the cooling coil inlet and outlet water temperature difference samples in the sliding window is taken, and then the square difference is divided by the sum of the square of the standard deviation of the cooling coil inlet and outlet water temperature difference and the constant term 1 to obtain the temperature difference deviation term; the temperature deviation term and the air pressure deviation term are added together. The flow deviation term, the temperature difference deviation term, and the flow deviation term are accumulated and the square root of the accumulated value is taken to obtain the multi-parameter normalized deviation sum, which is used to measure the degree of coordinated deviation between each sensor data and its short-term statistical characteristics. The difference between the current chip temperature and the target chip safety temperature is calculated, and the difference between the current box air pressure and the box saturated steam pressure is calculated. The two differences are multiplied and the absolute value is taken. The result is then divided by the product of the target chip safety temperature and the box saturated steam pressure to obtain the temperature-pressure interaction correction term, which is used to reflect the degree of coupling deviation between the heat load and the gas-liquid phase equilibrium state of the box. Among them, the target chip safety temperature is the reference temperature threshold in heat exchange control, representing the upper limit of the maximum temperature at which the chip can still operate stably under full load operation. It is used to determine whether the current chip temperature is within the safe range and guide the dynamic adjustment of the heat exchange control strategy. The temperature-pressure interaction correction term is added to the constant term of one to obtain the correction factor. Finally, the multi-parameter normalized deviation sum is multiplied by the correction factor to obtain the condenser operating condition abnormality assessment value corresponding to the sampling point, which is used to characterize the degree of deviation of the current condenser 3 thermal operation state from the chip heat exchange demand.

[0031] The specific calculation formula for the abnormal operating condition evaluation value of the condenser is: ; Where, S represents the abnormal operating condition assessment value of the condenser, Indicates the chip temperature, represents the average chip temperature, represents the chip temperature standard deviation, Indicates the box air pressure, Indicates the average box air pressure. Indicates the standard deviation of the box air pressure, Indicates the cooling water flow rate, represents the average cooling water flow rate, represents the standard deviation of cooling water flow, Indicates the water temperature difference between the inlet and outlet of the cooling coil, Indicates the average temperature difference between the inlet and outlet water of the cooling coil, It represents the standard deviation of the cooling coil inlet and outlet water temperature difference, Indicates the target chip safety temperature, Indicates the saturated steam pressure of the box.

[0032] This implementation effectively integrates the statistical deviation characteristics of four types of condenser heat exchange sensor data: chip temperature, box air pressure, cooling water flow rate, and the difference in cooling coil inlet and outlet water temperature. By introducing a coupling correction factor between the target chip safety temperature and the box saturated steam pressure, the robustness and sensitivity of anomaly identification are enhanced. This method enables accurate quantitative determination of abnormal operating conditions under different condenser operating states, improving the accuracy and dynamic adaptability of data screening and providing a reliable status basis for subsequent control instruction generation.

[0033] Specifically, the heat exchange adaptability of condenser 3 is evaluated based on the effective condenser operation data set, and the chip temperature deviation ratio and the box air pressure deviation ratio are extracted. The specific steps for identifying the load state category are as follows: extract the effective condenser operation data set retained after two rounds of data screening, and construct a parameter combination for evaluating the heat exchange capacity matching of condenser 3. Specifically, first, the three parameters of cooling water density, cooling water flow rate and cooling water specific heat capacity are multiplied by the water temperature difference between the inlet and outlet of the cooling coil to obtain the heat exchange power component reflecting the current heat exchange capacity of condenser 3; then, the difference between the box saturated steam pressure and the box air pressure is numerically compared with the minimum term, and the larger value is taken as the pressure margin. The degree component is used to maintain calculation stability. The minimum term is a non-zero, extremely small positive real number that provides a minimum positive limit when the difference between the saturated steam pressure of the box and the box air pressure approaches zero, preventing calculation anomalies caused by division by zero or numerical instability. The heat transfer power component is then divided by the pressure margin component to obtain the pressure-corrected heat transfer power that takes into account the pressure influencing factor. The difference between the chip temperature and the target chip safety temperature is then calculated and divided by the target chip safety temperature to obtain the standardized relative temperature deviation coefficient. The pressure-corrected heat transfer power is then multiplied by the relative temperature deviation coefficient to obtain the final condenser heat transfer adaptation assessment value that represents the degree of match between heat transfer capacity and load demand. The average chip temperature and box air pressure values ​​in the condenser heat transfer sensing data are further extracted and compared with the chip temperature threshold and box air pressure threshold, respectively. The chip temperature deviation ratio and box air pressure deviation ratio are calculated to reflect the relative distance between the current heat load state and the safety boundary. The condenser heat exchange adaptation evaluation value, the chip temperature deviation ratio and the box air pressure deviation ratio are combined to form a feature vector, which is input into the BP neural network classifier, and the load state category of the current condenser 3 operation is output. The load state category includes high load state, medium load state and low load state, which is used to provide a key decision basis for the subsequent dynamic control instruction generation.

[0034] The specific calculation formula for the condenser heat exchange adaptation evaluation value is: ; Where H represents the condenser heat transfer adaptation evaluation value, represents cooling water density, C represents cooling water specific heat capacity, Q represents cooling water flow rate, Indicates the water temperature difference between the inlet and outlet of the cooling coil, Indicates the saturated steam pressure of the box. Indicates the box air pressure, Indicates the chip temperature, Indicates the target chip safety temperature, Indicates a minimum term.

[0035] In this embodiment, Table 1 is a table of condenser heat exchange adaptation evaluation values, which lists the cooling water density, cooling water specific heat capacity, cooling water flow rate, cooling coil inlet and outlet water temperature difference, box saturated steam pressure, box air pressure, chip temperature, target chip safety temperature at 5 sampling points, and the condenser heat exchange adaptation evaluation value calculated based on the above variables. The specific data are as follows: At sampling point 1, the cooling water density is 995.99, the cooling water specific heat capacity is 4081.41, the cooling water flow rate is 0.03959, the cooling coil inlet and outlet water temperature difference is 2.72, the box saturated steam pressure is 105.48, the box air pressure is 100.8, the chip temperature is 66.15, the target chip safety temperature is 66.1, and the calculated condenser heat exchange adaptation evaluation value is 78.51; at sampling point 2, the cooling water density is 1058.58, the cooling water specific heat capacity is 40 89.35, cooling water flow rate is 0.03837, cooling coil inlet and outlet water temperature difference is 2.36, box saturated steam pressure is 115.59, box air pressure is 111.54, chip temperature is 68.69, target chip safety temperature is 68.63, the calculated condenser heat exchange adaptation evaluation value is 95.51; at sampling point 3, the cooling water density is 935.82, the cooling water specific heat capacity is 4137.77, the cooling water flow rate is 0.03551, the cooling coil inlet and outlet water temperature difference is 4.43, box The saturated steam pressure of the body is 113.69, the box air pressure is 109.81, the chip temperature is 65.71, the target chip safety temperature is 65.68, and the calculated condenser heat exchange adaptation evaluation value is 73.08; at sampling point 4, the cooling water density is 950.60, the cooling water specific heat capacity is 4046.01, the cooling water flow rate is 0.04457, the cooling coil inlet and outlet water temperature difference is 2.26, the box saturated steam pressure is 118.51, the box air pressure is 113.8, and the chip temperature is 65.12 , the target chip safety temperature is 65.06, and the calculated condenser heat exchange adaptation evaluation value is 69.39; at sampling point 5, the cooling water density is 926.49, the cooling water specific heat capacity is 4160.34, the cooling water flow rate is 0.04099, the cooling coil inlet and outlet water temperature difference is 2.20, the box saturated steam pressure is 113.31, the box air pressure is 110.71, the chip temperature is 68.81, the target chip safety temperature is 68.77, and the calculated condenser heat exchange adaptation evaluation value is 73.91.

[0036] Table 1 Condenser heat transfer adaptation evaluation value data table

[0037] like Figure 3The figure shows the changing trends of the condenser heat transfer adaptation evaluation values ​​at five different sampling points. The horizontal axis represents the sampling point number, and the vertical axis represents the corresponding heat transfer adaptation evaluation value, which is used to measure the degree of match between the current operating condition of Condenser 3 and the target heat transfer state. As can be seen from the figure: the evaluation value at sampling point 2 is the highest, indicating that Condenser 3 has the optimal heat transfer capacity at this moment; the evaluation value at sampling point 4 is the lowest, indicating relatively weak heat transfer efficiency, which may require strategic intervention; the overall evaluation value remains within the range of [69,96], indicating that the operating condition of Condenser 3 fluctuates relatively mildly. The data labels clearly identify the value of each point, facilitating the analysis of fluctuation patterns and trend judgment. Figure 3 It can be used to intuitively evaluate the execution effect of the heat exchange control strategy of condenser 3 and the adaptability of the system operation status. It is an important visualization support for reinforcement learning strategy optimization and fault warning.

[0038] In this implementation scheme, a calculation model coupling multiple parameters such as cooling water density, cooling water flow, cooling water specific heat capacity, cooling coil inlet and outlet water temperature difference, box saturated steam pressure, box air pressure, chip temperature and target chip safety temperature is constructed to quantitatively evaluate the condenser heat exchange adaptation evaluation value, and combined with the chip temperature deviation ratio and the box air pressure deviation ratio, a state feature vector is formed that can comprehensively reflect the degree of matching between the heat exchange capacity and the current heat load. The vector is input into the BP neural network classifier to realize accurate identification of the load state category, improve the accuracy of the state judgment before dynamic regulation of condenser 3, and provide a high-reliability decision-making basis for subsequent reinforcement learning strategy generation.

[0039] Specifically, the heat exchange adaptation evaluation results, load state category and effective condenser operation data are input into the reinforcement learning decision model, and the specific steps for outputting the control instructions are as follows: based on the historical effective condenser operation data, a sequence of key parameters including cooling water density, cooling water flow, cooling water specific heat capacity, cooling coil inlet and outlet water temperature difference, chip temperature, box air pressure and pump power are collected to build a long-term operation state sequence database, and the Q-Learning algorithm is used to build a reinforcement learning decision model. In the model initialization stage, the target cooling coil inlet water temperature adjustment amount and the target pump power adjustment amount are used as discrete control variables in the action set. The load state category, condenser heat exchange adaptation evaluation value and effective condenser operation data including cooling coil inlet water temperature, cooling coil outlet water temperature, cooling water density, cooling water flow, cooling water specific heat capacity, pump power, chip temperature and box air pressure are set as high-dimensional input features in the state set. The compound reward function is to maximize the heat dissipation efficiency under high load and minimize the unit energy consumption under low load. Iterative training is performed to complete the reinforcement learning decision model training stage. During the policy inference phase, the current load state category, condenser heat transfer adaptation evaluation value, and all valid condenser operating data are input as decision input parameters into the reinforcement learning decision model, which then outputs corresponding control instructions: Under high load conditions, the cold target cooling coil inlet water temperature is actively reduced while the target pump power is simultaneously increased to enhance the heat transfer intensity of Condenser 3; under low load conditions, the cold target cooling coil inlet water temperature is increased while the target pump power is reduced to reduce cooling system energy consumption; under medium load conditions, the current cold target cooling coil inlet water temperature and target pump power are maintained to maintain stable heat transfer efficiency. The target cooling coil inlet water temperature is the ideal cooling water inlet temperature derived from the current load state and heat transfer demand. It is used to guide the cooling system to dynamically adjust the cooling intensity to ensure a balance between chip heat dissipation demand and energy consumption, and serves as the key control variable for the control instructions. The target pump power is a pump operating power setpoint calculated based on the operating state of Condenser 3, the heat transfer adaptation level, and the energy consumption optimization goal. It is used to adjust the cooling water circulation rate and control the overall energy efficiency of the system. It is compared with the actual power in the feedback evaluation to guide subsequent optimization of the control strategy.

[0040] In this implementation scheme, by jointly inputting the heat exchange adaptation evaluation results of condenser 3, the load state category and the effective condenser operation data into the reinforcement learning decision model, the dynamic output of the cold target cooling coil inlet water temperature adjustment amount and the target pump power adjustment amount is achieved, and a closed-loop control mechanism of perception-decision-execution is constructed. This mechanism is based on the Q-Learning algorithm, which fully exploits the behavior patterns of key variables such as cooling water density, cooling water flow, cooling water specific heat capacity, cooling coil inlet and outlet water temperature difference, chip temperature, cabinet air pressure and pump power contained in the historical effective condenser operation data. With the maximization of heat dissipation efficiency under high load conditions and the minimization of energy consumption under low load conditions as the reward orientation, a decision model with optimal strategy convergence capability is trained. Compared with the traditional fixed control logic that relies on experience-based settings, it realizes the adaptive control output for the operating conditions of condenser 3 under different load conditions, significantly improving the operating energy efficiency of condenser 3 and the flexibility of the control strategy.

[0041] Specifically, the steps for executing a control command and collecting execution feedback data are as follows: receiving a control command, sending the target cooling coil inlet water temperature adjustment value to the cooling water temperature control execution unit, which drives the cooling water temperature control execution unit to adjust the temperature based on the target value, thereby achieving dynamic control of the cooling coil inlet water temperature; sending the target pump power adjustment value to the pump speed control unit, which causes the pump speed control unit to adjust the pump operating state based on the target power, thereby achieving precise regulation of the pump power output. During the execution of the control command, the actual values ​​of the cooling coil inlet water temperature, the cooling coil outlet water temperature, and the pump power after execution are collected in real time. The start and completion times of the command response are also recorded to calculate the actual execution response time. During the control cycle, energy consumption data recorded by the energy metering unit is synchronously obtained as the actual total energy consumption. An execution feedback dataset is constructed based on the actual values ​​of the cooling coil inlet water temperature, the actual values ​​of the cooling coil outlet water temperature, the actual pump power, the actual execution response time, and the actual total energy consumption, providing a complete data foundation for the subsequent evaluation of the heat exchange execution performance of condenser 3.

[0042] In this embodiment, the target cooling coil inlet water temperature adjustment amount is accurately sent to the cooling water temperature control execution unit, and the target pump power adjustment amount is simultaneously sent to the pump speed control unit to ensure efficient response of the control instructions at the cooling coil inlet water temperature and pump power levels. During the control execution process, the actual value of the cooling coil inlet water temperature, the actual value of the cooling coil outlet water temperature, the actual value of the pump power, the actual execution response time and the actual total energy consumption during the control cycle are collected in real time to construct an execution feedback data set. This feedback data set constitutes a necessary input source for the subsequent evaluation of the heat exchange execution performance of condenser 3, provides a basis for dynamic analysis of the cooling coil inlet water temperature regulation effect and pump power control accuracy, and lays the foundation for judging the execution quality of the cooling control strategy, thereby improving the thermal management adaptability of condenser 3 under variable load conditions.

[0043] Specifically, the specific steps for evaluating the heat exchange execution performance of condenser 3 are as follows: based on the execution feedback data set, the values ​​of cooling water density, cooling water flow and cooling water specific heat capacity are extracted, and the cooling water density is multiplied by the cooling water flow and then multiplied by the cooling water specific heat capacity in turn to calculate the total value of the cooling water heat capacity flow, and then multiplied by the temperature difference between the actual value of the cooling coil inlet water temperature after execution and the actual value of the cooling coil outlet water temperature after execution to obtain the heat exchange power component; the actual execution response time is calculated by ratio with the benchmark response time, and one is added to the ratio to obtain the response time correction term, which is used to reflect the dynamic delay degree of the control response process; wherein, the benchmark response time is a response time reference value, which represents the ideal time required for the actuator to complete the control instruction under standard working conditions, and is used to measure whether the current control response is lagging or over-delayed. Fast, and as an important parameter for evaluating execution performance, it guides the dynamic correction of the execution speed of the control strategy; the absolute value of the difference between the actual value of the cooling coil inlet water temperature after execution and the target cooling coil inlet water temperature, as well as the absolute value of the difference between the actual value of the pump power after execution and the target pump power are calculated respectively, and the two absolute differences are added together and added with one to obtain the execution deviation correction term, which is used to measure the joint deviation degree of the cooling water temperature control accuracy and the pump power control accuracy; the actual total energy consumption, the response time correction term and the execution deviation correction term in the control cycle are multiplied to calculate the total corrected energy consumption, which is used to comprehensively reflect the comprehensive energy efficiency performance in the control process; finally, the heat exchange power component is divided by the total corrected energy consumption to obtain the heat exchange execution performance evaluation value of the current control cycle of condenser 3, which is used as an important performance indicator for feedback of the effect of the control strategy.

[0044] The specific calculation formula for the heat exchange performance evaluation value is: ; Where X represents the heat transfer performance evaluation value, represents cooling water density, Q represents cooling water flow rate, C represents cooling water specific heat capacity, Indicates the actual value of the cooling coil inlet water temperature after execution. Indicates the actual value of the cooling coil outlet water temperature after execution, W indicates the actual total energy consumption during the control cycle, Indicates the actual execution response time, represents the benchmark response time, Indicates the target cooling coil inlet water temperature, G indicates the actual pump power value after execution, Indicates the target pump power.

[0045] In this implementation scheme, a heat exchange execution performance evaluation mechanism is constructed with cooling water density, cooling water flow rate, cooling water specific heat capacity, actual value of cooling coil inlet water temperature after execution, actual value of cooling coil outlet water temperature after execution, actual execution response time, benchmark response time, target cooling coil inlet water temperature, actual value of cooling coil inlet water temperature after execution, target pump power, actual value of pump power after execution, and actual total energy consumption during the control cycle as core variables, effectively achieving dynamic performance quantitative analysis of the control execution process. This method comprehensively evaluates the control response efficiency and execution accuracy by calculating the heat exchange power component, response time correction term, and execution deviation correction term, and uses the total corrected energy consumption as the energy efficiency benchmark to ultimately obtain the heat exchange execution performance evaluation value, providing a quantifiable, feedback-based, and closed-loop evaluation basis for the effectiveness judgment and optimization of subsequent control strategies, significantly enhancing the accuracy and adaptability of control decisions.

[0046] Specifically, based on the evaluation results, it is determined whether the current control instructions need to be adjusted, and the execution feedback data and evaluation results are sent back. The specific steps for threshold update and control strategy optimization are as follows: real-time comparison of the heat exchange execution performance evaluation value and the execution performance threshold. After each control cycle, the current heat exchange execution performance evaluation value is judged and analyzed according to the performance threshold standard; when the heat exchange execution performance evaluation value is greater than or equal to the execution performance threshold, it is determined that the current control strategy meets the operating requirements in terms of control accuracy, response speed and energy efficiency, and the current control strategy is maintained to enter the next control cycle for continued execution; when the heat exchange execution performance evaluation value is less than the execution performance threshold, based on the absolute difference between the actual value of the cooling coil inlet water temperature after execution and the target cooling coil inlet water temperature recorded in the execution feedback data set, and the absolute difference between the actual value of the pump power after execution and the target pump power, the cooling coil inlet water temperature deviation and pump power are constructed. rate deviation; according to the sign characteristics and numerical amplitude of the cooling coil inlet water temperature deviation and the pump power deviation, the current control offset direction and correction degree are calculated to generate the corrected cold target cooling coil inlet water temperature adjustment amount and the corrected target pump power adjustment amount, and the above control parameters are combined into a correction control instruction and sent to the actuator to drive the cooling water temperature control execution unit and the pump speed control unit to perform correction adjustment; regardless of whether the strategy correction is performed, the execution feedback data set collected in the current cycle and the calculated heat exchange execution performance evaluation value are synchronously transmitted back to the data screening link and the intelligent decision-making link, and the chip temperature threshold, box air pressure threshold and cooling water flow threshold are corrected in real time in the data screening link, and the state behavior mapping relationship corresponding to the cold target cooling coil inlet water temperature adjustment strategy and the target pump power adjustment strategy is updated in the intelligent decision-making link, so as to realize the multi-dimensional dynamic heat exchange control closed-loop operation mechanism for the operating state of condenser 3.

[0047] In this implementation scheme, based on the real-time comparison between the heat exchange execution performance evaluation value and the execution performance threshold, a dynamic judgment and correction mechanism of the control instruction effect is realized to ensure the accuracy of the control strategy in the cooling coil inlet water temperature adjustment and the target pump power setting; by constructing a corrected control instruction based on the cooling coil inlet water temperature deviation and the pump power deviation in the execution feedback data set, and feeding the execution feedback data and the heat exchange execution performance evaluation value back to the data screening link and the intelligent decision-making link, they are respectively used to update the chip temperature threshold, the box air pressure threshold, the cooling water flow threshold and the optimization of the cold target cooling coil inlet water temperature adjustment strategy and the target pump power adjustment strategy, so as to realize continuous adaptive correction and precision enhancement of the control closed loop, and improve the stability, robustness and energy efficiency performance of the heat exchange control.

[0048] like Figure 2 As shown, the second aspect of the present invention provides a condenser dynamic heat exchange control system based on multi-dimensional perception and AI, including: a data perception preprocessing module, an abnormal screening emergency control module, an intelligent state judgment decision generation module and an execution feedback closed-loop optimization module, wherein: the data perception preprocessing module is used to collect condenser heat exchange perception data, and perform time series synchronization, noise separation, physical consistency verification and normalization mapping processing on the condenser heat exchange perception data; the abnormal screening emergency control module is used to perform two rounds of data screening on the sampling points based on the preprocessed condenser heat exchange perception data, realize the issuance of emergency control instructions and abnormal point data processing, and build an effective condenser. Condenser operation data set; intelligent state judgment decision generation module, used to evaluate the heat exchange adaptation degree of condenser 3 based on the effective condenser operation data set, and extract the chip temperature deviation ratio and the box air pressure deviation ratio to identify the load state category; input the heat exchange adaptation evaluation results, load state category and effective condenser operation data into the reinforcement learning decision model, and output the control instructions; execution feedback closed-loop optimization module, used to execute the control instructions and collect execution feedback data, evaluate the heat exchange execution performance of condenser 3, determine whether the current control instructions need to be adjusted based on the evaluation results, and send back the execution feedback data and evaluation results for threshold update and control strategy optimization.

[0049] like Figure 4As shown, saturated vapor pressure sensor 1 is used to collect the saturated vapor pressure of the cabinet, forming an important component of the condenser heat exchange sensing data. Cooling coil 2 is installed inside condenser 3, performing heat exchange between the coolant inlet and outlet, regulating the heat exchange efficiency of condenser 3. Condenser 3 is located above the heat dissipation path of server 4. When the chips generate steam during operation, it promptly condenses the steam into liquid, effectively releasing heat. Server 4 is the primary heat source, and its chips continuously release heat during operation, creating a demand for heat exchange with condenser 3. Temperature sensors 5 are deployed at the inlet and outlet of cooling coil 2, respectively, collecting the cooling coil inlet and outlet water temperatures for calculating the cooling coil inlet and outlet water temperature difference. Pump 6 drives the continuous circulation of coolant and adjusts pump power according to control instructions to achieve precise control of the target pump power. Coolant reservoir 7 stores coolant to ensure liquid supply and circulation stability in the cooling circuit.

[0050] In this implementation scheme, by constructing a condenser dynamic heat exchange control system including a data perception preprocessing module, an abnormality screening emergency control module, an intelligent state judgment decision generation module and an execution feedback closed-loop optimization module, the whole process of condenser heat exchange perception data processing, abnormality identification and emergency response, operating status intelligent identification and control instruction generation, execution effect feedback evaluation and strategy continuous optimization is systematically covered; the functional boundaries of each module are clear, and the condenser heat exchange perception data enters the data screening mechanism after time synchronization, noise separation, physical consistency verification and normalization mapping processing to form an effective condenser operation data set, and then through load state identification and heat exchange adaptation evaluation, drives the reinforcement learning decision model to generate control instructions, and finally through heat exchange execution performance evaluation and execution feedback data transmission, realizes continuous optimization of the control strategy, and ensures the accuracy, adaptability and closed-loop robustness of the dynamic heat exchange control of condenser 3.

[0051] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0052] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A condenser dynamic heat exchange control method based on multi-dimensional perception and AI, characterized by: The following steps are involved: S1, collect condenser heat exchange perception data, perform time series synchronization, noise separation, physical consistency check and normalization mapping processing on the condenser heat exchange perception data; S2, based on the pre-processed condenser heat exchange sensing data, performs two rounds of data screening on the sampling points to implement emergency control instructions and abnormal point data processing, and construct an effective condenser operation data set; S3, based on the effective condenser operation data set, evaluates the heat exchange adaptation degree of the condenser (3), extracts the chip temperature deviation ratio and the box air pressure deviation ratio, and identifies the load state category; inputs the heat exchange adaptation evaluation results, load state category and effective condenser operation data into the reinforcement learning decision model, and outputs the control instructions; S4, executes the control instruction and collects the execution feedback data, evaluates the heat exchange execution performance of the condenser (3), determines whether the current control instruction needs to be adjusted based on the evaluation result, and transmits the execution feedback data and the evaluation result back for threshold update and control strategy optimization.

2. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 1 is characterized by: The specific steps of collecting the condenser heat exchange perception data and performing time series synchronization, noise separation, physical consistency verification and normalization mapping processing on the condenser heat exchange perception data are as follows: Collect condenser heat exchange sensing data, including chip temperature, box air pressure, cooling coil inlet water temperature, cooling coil outlet water temperature, cooling water density, cooling water flow, cooling water specific heat capacity, pump power, and box saturated steam pressure; Through multi-channel timestamp comparison and sampling delay compensation methods, the condenser heat exchange perception data is synchronized across sensors and the sampling rate is unified; through a multi-scale signal reconstruction method combining Kalman filtering and wavelet decomposition, the measurement noise of the condenser heat exchange perception data is eliminated and the periodic disturbance is separated; based on the energy conservation equation and the pressure difference flow coupling model, a heat flow integrated consistency verification mechanism is established to identify and eliminate data that does not conform to physical laws in the condenser heat exchange perception data; through multi-dimensional feature encoding and transmission protocol adaptive encapsulation methods, the data format of the condenser heat exchange perception data is standardized and the interface compatibility conversion is processed; through physical quantity normalization mapping and dynamic interval stretching algorithm, the unit conversion and feature scale normalization of the condenser heat exchange perception data are carried out.

3. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 1 is characterized by: Based on the pre-processed condenser heat exchange sensing data, two rounds of data screening are performed on the sampling points to implement the issuance of emergency control instructions and abnormal point data processing. The specific steps for constructing an effective condenser operation data set are as follows: Extract the pre-processed condenser heat exchange sensing data and perform the first round of data screening for each sampling point: when the chip temperature exceeds the chip temperature upper threshold, the box air pressure exceeds the air pressure upper threshold, or the cooling water flow rate is lower than the flow lower threshold, the data is judged to be abnormal and emergency control is triggered. The target cooling coil inlet water temperature and target pump power are adjusted to the emergency set value and sent to the actuator. The remaining sampling points enter the second round of screening; For the sampling points that enter the second screening, the cooling coil inlet and outlet water temperature difference is calculated based on the difference between the cooling coil inlet water temperature and the cooling coil outlet water temperature. The mean and standard deviation of the chip temperature, box air pressure, cooling water flow rate and cooling coil inlet and outlet water temperature difference are calculated within the sliding window to evaluate the abnormality of the condenser (3) working condition and obtain the abnormality evaluation value of the condenser working condition. Compare the condenser operating condition abnormality assessment value and the operating condition abnormality threshold in real time. When the condenser operating condition abnormality assessment value exceeds the abnormality scoring threshold, the corresponding sampling point is determined to be an abnormal point. If it is an isolated abnormal point, it is replaced by the sliding average of the condenser heat exchange sensing data of the adjacent sampling points. If three or more consecutive sampling points are determined to be abnormal points and the corresponding parameter collection source is the same sensor, the backup sensor collection value is switched. Otherwise, the corresponding sampling point is determined to be a normal point and the original data is retained. The condenser heat exchange perception data retained after two rounds of data screening are used to construct an effective condenser operation data set in the order of sampling points.

4. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 3 is characterized by: The specific steps of evaluating the abnormality of the condenser operating condition and obtaining the abnormality evaluation value of the condenser operating condition are as follows: The temperature deviation term is obtained by dividing the square of the difference between the chip temperature and the chip temperature mean by the square of the chip temperature standard deviation plus one. The pressure deviation term is obtained by dividing the square of the difference between the box air pressure and the box air pressure mean by the square of the box air pressure standard deviation plus one. The flow deviation term is obtained by dividing the square of the difference between the cooling water flow rate and the cooling water flow mean by the square of the cooling water flow standard deviation plus one. The temperature difference deviation term is obtained by dividing the square of the difference between the inlet and outlet water temperature of the cooling coil and the mean by the square of the temperature difference standard deviation plus one. The temperature deviation term, air pressure deviation term, flow deviation term, and temperature difference deviation term are added together and then squared to obtain the multi-parameter normalized deviation sum; the difference between the chip temperature and the target chip safety temperature is multiplied by the difference between the box air pressure and the box saturated steam pressure, the absolute value is taken and divided by the product of the target chip safety temperature and the box saturated steam pressure to obtain the temperature-pressure interaction correction term; the temperature-pressure interaction correction term is added by one to obtain the correction factor; the multi-parameter normalized deviation sum is multiplied by the correction factor to obtain the condenser operating condition abnormality assessment value.

5. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 1 is characterized by: The specific steps of evaluating the heat exchange adaptability of the condenser (3) based on the effective condenser operation data set, extracting the chip temperature deviation ratio and the box air pressure deviation ratio, and identifying the load state category are as follows: Extract the effective condenser operation data set and evaluate the heat transfer adaptation degree of the condenser (3): multiply the cooling water density, cooling water flow rate and cooling water specific heat capacity, and then multiply by the water temperature difference between the inlet and outlet of the cooling coil to obtain the heat transfer power component; take the larger value of the difference between the box saturated steam pressure and the box air pressure and the minimum term to obtain the pressure margin component; divide the heat transfer power component by the pressure margin component to obtain the pressure-corrected heat transfer power; divide the difference between the chip temperature and the target chip safety temperature by the target chip safety temperature to obtain the relative temperature deviation coefficient; multiply the pressure-corrected heat transfer power by the relative temperature deviation coefficient to obtain the condenser heat transfer adaptation evaluation value; Extract the average chip temperature and the average box pressure, compare them with the chip temperature threshold and the box pressure threshold respectively, and obtain the chip temperature deviation ratio and the box pressure deviation ratio; The condenser heat exchange adaptation evaluation value, chip temperature deviation ratio and box air pressure deviation ratio are combined into a feature vector, which is input into the BP neural network classifier to output the current load state category, including high load, medium load and low load.

6. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 1 is characterized by: The specific steps of inputting the heat exchange adaptation evaluation results, load state categories, and effective condenser operation data into the reinforcement learning decision model and outputting the control instructions are as follows: Based on the historical effective condenser operation data, the Q-Learning algorithm is used to construct a reinforcement learning decision model. The cold target cooling coil inlet water temperature adjustment amount and the target pump power adjustment amount are set as the action set, and the load state category, condenser heat exchange adaptation evaluation value and effective condenser operation data are set as the state set. The reinforcement learning decision model training is completed with maximizing the heat dissipation efficiency under high load and minimizing the energy consumption under low load as the reward function; the current load state category, condenser heat exchange adaptation evaluation value and effective condenser operation data are input as decision input parameters into the reinforcement learning decision model, and the control instructions are output: reduce the cold target cooling coil inlet water temperature and increase the target pump power under high load; increase the cold target cooling coil inlet water temperature and reduce the target pump power under low load; maintain the current cold target cooling coil inlet water temperature and target pump power under medium load.

7. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 1 is characterized by: The specific steps of executing the control instruction and collecting execution feedback data are as follows: Receive the control instructions, send the target cooling coil inlet water temperature adjustment amount to the cooling water temperature control execution unit, send the target pump power adjustment amount to the pump speed control unit, and drive the control execution for real-time adjustment; during the control execution process, collect the actual value of the cooling coil inlet water temperature after execution, the actual value of the cooling coil outlet water temperature after execution, the actual value of the pump power after execution, the actual execution response time and the actual total energy consumption during the control cycle to construct an execution feedback data set.

8. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 1 is characterized by: The specific steps of evaluating the condenser heat exchange performance are as follows: Based on the execution feedback data set, the cooling water density, cooling water flow rate and cooling water specific heat capacity are multiplied together, and then multiplied by the difference between the actual value of the cooling coil inlet water temperature after execution and the actual value of the cooling coil outlet water temperature after execution to obtain the heat exchange power component; the actual execution response time is divided by the benchmark response time and added one to obtain the response time correction term; the absolute value of the difference between the actual value of the cooling coil inlet water temperature after execution and the target cooling coil inlet water temperature is calculated, and the absolute value of the difference between the actual value of the pump power after execution and the target pump power is calculated. The two absolute values ​​are added together and then added one to obtain the execution deviation correction term; the actual total energy consumption, response time correction term and execution deviation correction term in the control period are multiplied together to obtain the total corrected energy consumption; the heat exchange power component is divided by the total corrected energy consumption to obtain the heat exchange execution performance evaluation value.

9. The condenser dynamic heat exchange control method based on multi-dimensional perception and AI according to claim 1 is characterized by: The specific steps of determining whether to adjust the current control instructions based on the evaluation results and transmitting the execution feedback data and the evaluation results back for threshold update and control strategy optimization are as follows: The heat exchange execution performance evaluation value is compared with the execution performance threshold in real time. When the heat exchange execution performance evaluation value is greater than or equal to the execution performance threshold, the current control strategy is maintained to execute the next control cycle; when the heat exchange execution performance evaluation value is less than the execution performance threshold, based on the correspondence between the actual value of the cooling coil inlet water temperature after execution and the target cooling coil inlet water temperature, and the actual value of the pump power after execution and the target pump power recorded in the execution feedback data set, the cooling coil inlet water temperature deviation and pump power deviation are determined, and according to the direction and amplitude of the deviation, the corrected cold target cooling coil inlet water temperature and target pump power instructions are generated and sent to the execution agency; regardless of whether the strategy is corrected, the execution feedback data set and the heat exchange execution performance evaluation value are respectively transmitted back to the data screening link and the intelligent decision-making link, the temperature threshold, air pressure threshold and flow threshold are updated in the data screening link, and the cold target cooling coil inlet water temperature adjustment strategy and the target pump power adjustment strategy are optimized in the intelligent decision-making link to realize the closed-loop operation of dynamic heat exchange control.

10. The condenser dynamic heat exchange control system based on multi-dimensional perception and AI is characterized by: include: Data perception preprocessing module, abnormal screening emergency control module, intelligent state judgment decision generation module and execution feedback closed-loop optimization module, among which: The data perception preprocessing module is used to collect condenser heat exchange perception data and perform time sequence synchronization, noise separation, physical consistency verification and normalization mapping processing on the condenser heat exchange perception data; The abnormal screening emergency control module is used to perform two rounds of data screening on the sampling points based on the pre-processed condenser heat exchange perception data, implement emergency control instructions and abnormal point data processing, and build an effective condenser operation data set; The intelligent state judgment decision generation module is used to evaluate the heat exchange adaptation degree of the condenser (3) based on the effective condenser operation data set, and extract the chip temperature deviation ratio and the box air pressure deviation ratio to identify the load state category; the heat exchange adaptation evaluation result, the load state category and the effective condenser operation data are input into the reinforcement learning decision model, and the control instruction is output; The execution feedback closed-loop optimization module is used to execute the control instructions and collect execution feedback data, evaluate the heat exchange execution performance of the condenser (3), determine whether the current control instructions need to be adjusted based on the evaluation results, and return the execution feedback data and the evaluation results for threshold update and control strategy optimization.

Images