Liquid cooling radiator cooling liquid leakage detection method and system

Abstract

The invention discloses a liquid cooling radiator cooling liquid leakage detection method and system, and relates to the technical field of equipment detection, and the method comprises the steps: obtaining the pressure reading of pressure measurement equipment; according to the pressure reading, the pressure change rate is calculated through differential operation; identifying a starting time point of a leakage event according to the pressure change rate; calculating an inherent response delay of the pressure measurement device; compensating the occurrence time sequence of the leakage event according to the inherent response delay to obtain a compensation time sequence; according to the starting time point and the compensation time sequence, the target leakage point position is determined, cooling liquid leakage detection is achieved, and accuracy is improved.

Description

Technical Field

[0001] This invention relates to the field of equipment testing technology, and in particular to a method and system for detecting coolant leakage in liquid-cooled radiators. Background Technology

[0002] In large data centers, liquid cooling infrastructure is widely deployed to ensure the stable operation of high-density computing equipment. This system typically consists of multiple liquid coolers connected in series to form a closed loop, with coolant flowing within it driven by a circulation pump. To monitor system status in real time, high-precision pressure measuring devices are installed at key locations in the piping to continuously collect pressure data and transmit it to the central control unit. However, new coolant additives introduced during actual operation may unexpectedly trigger a polymerization reaction with existing corrosion inhibitors under the high-speed shearing action of the coolant circulation pump, generating micron-sized colloidal suspended particles. Due to their size characteristics, these particles cannot be effectively intercepted by existing filtration devices, potentially leading to a slight increase in the overall viscosity of the coolant, increasing flow resistance, and subtly affecting the system pressure distribution. Existing methods establish a calibration parameter set under stable operating conditions with clean coolant, assuming all pressure measuring devices have consistent response characteristics and that their readings are free from any offset or delay. However, with the occurrence of the aforementioned particle deposition, a significant deviation emerges between the underlying data model upon which the system operation relies and the actual physical reality, resulting in detection errors and low accuracy.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The main objective of this invention is to propose a method and system for detecting coolant leakage in liquid-cooled radiators, which can combine pressure change rate and inherent response delay to detect coolant leakage, thereby improving accuracy.

[0005] On one hand, embodiments of the present invention provide a method for detecting coolant leakage in a liquid-cooled radiator, comprising the following steps: Obtain pressure readings from pressure measuring equipment; Based on the pressure readings, the rate of pressure change is calculated using differential calculations. Based on the pressure change rate, identify the start time of the leak event; Calculate the inherent response delay of the pressure measuring device; Based on the inherent response delay, the timing of the leakage event is compensated to obtain the compensated timing. The location of the target leak point is determined based on the starting time point and the compensation sequence.

[0006] In some embodiments, calculating the pressure change rate using differential operations based on the pressure reading includes: Obtain historical stress data; The historical pressure data and the pressure readings are subjected to feature analysis to obtain fluctuation characteristics, which include fluctuation amplitude or frequency distribution. Based on the fluctuation characteristics, the time interval for differential operations is determined; The pressure change rate is calculated based on the historical pressure data, the pressure readings, and the differential calculation time interval.

[0007] In some embodiments, identifying the start time of the leak event based on the rate of pressure change includes: If the rate of pressure change is greater than a preset threshold, a leakage signal is obtained; Local waveform analysis was performed on the leakage signal to obtain the local waveform of the signal; Within a preset time window, target points in the local waveform of the signal are identified, including waveform inflection points or local extreme points. The timestamp corresponding to the target point is used as the starting time point.

[0008] In some embodiments, calculating the inherent response delay of the pressure measuring device includes: Acquire reference pressure waveform data segments and pressure waveform data streams to be matched; Perform cross-correlation calculation on the reference pressure waveform data segment and the pressure waveform data stream to be matched to obtain the cross-correlation function; The actual propagation time difference is determined based on the time offset corresponding to the peak point with the largest amplitude in the cross-correlation function. Calculate the theoretical propagation time difference between adjacent pressure measuring devices based on the physical distance between them and the sound velocity of the coolant. The inherent response delay is calculated based on the actual propagation time difference and the theoretical propagation time difference.

[0009] In some embodiments, determining the location of the target leak point based on the start time point and the compensation timing includes: Obtain the physical properties of the coolant in the cooling pipes; Calculate the speed of sound of the coolant in the cooling pipes based on the physical properties of the coolant. Set multiple candidate leak point locations; Based on the sound velocity of the coolant, calculate the theoretical propagation time of the pressure wave from the candidate leak point to the pressure measuring device; Based on the starting time point and the pressure change rate, the location of the first leak point is identified; Based on the theoretical propagation time and the compensation timing, identify the location of the second leakage point; The location of the target leak point is determined based on the location of the first leak point and the location of the second leak point.

[0010] In some embodiments, identifying the location of the first leak point based on the starting time point and the pressure change rate includes: The absolute value of the rate of pressure change is taken as the event intensity; Spatiotemporal correlation analysis was performed on multiple starting time points to obtain multiple pressure wave propagation paths; Identify the trend of event intensity variation in each of the pressure wave propagation paths; If the trend of change is downward, then the location of the first leak point is identified based on the starting time point and the laws of fluid physics propagation.

[0011] In some embodiments, obtaining the physical properties of the coolant in the cooling pipes includes: The cooling pipes are divided into multiple pipe sections; Collect physical property data for each pipe segment; Calculate the physical properties of the coolant in a single pipe section based on the fluid path, heat exchange characteristics, and physical property data of the pipe section. The physical properties of the coolant in the cooling pipeline are obtained by performing time-smoothing processing on the physical properties of the coolant in multiple single-pipe sections.

[0012] In some embodiments, calculating the physical properties of the coolant in a single pipe segment based on the fluid path, heat exchange characteristics, and physical property data of the pipe segment includes: The physical property data are then filtered. Based on the filtered physical property data and the energy and mass conservation rules of the pipe section, the target parameters are verified. The target parameters include the parameters of the fluid path or the parameters of the heat exchange characteristics. Based on the verified target parameters, calculate the physical properties of the coolant in the single pipe section.

[0013] In some embodiments, calculating the physical properties of the coolant in a single pipe section based on the verified target parameters includes: Obtain the thermodynamic property parameters of the coolant and the heat transfer coefficient of the pipe section; The physical properties of the coolant in a single pipe section are calculated based on the thermodynamic property parameter table of the coolant, the heat transfer coefficient of the pipe section, and the verified target parameters.

[0014] On the other hand, embodiments of the present invention provide a liquid coolant leakage detection system for liquid-cooled radiators, comprising: The pressure data acquisition module is used to acquire pressure readings from pressure measuring equipment. The rate calculation module is used to calculate the rate of pressure change based on the pressure reading using differential operations; A time recording module is used to identify the start time of a leakage event based on the pressure change rate; The response delay quantization module is used to calculate the inherent response delay of the pressure measuring device; The timing compensation module is used to compensate for the occurrence timing of the leakage event based on the inherent response delay, so as to obtain the compensated timing. The leak location module is used to determine the location of the target leak point based on the start time point and the compensation timing.

[0015] The embodiments of this application include at least the following beneficial effects: First, the pressure reading of the pressure measuring device is obtained. Based on the pressure reading, the pressure change rate is calculated using differential calculation. Then, based on the pressure change rate, the start time point of the leakage event is identified. Next, the inherent response delay of the pressure measuring device is calculated. Based on the inherent response delay, the occurrence sequence of the leakage event is compensated to obtain the compensated sequence. Finally, based on the start time point and the compensated sequence, the location of the target leakage point is determined. Thus, the coolant leakage detection can be achieved by combining the pressure change rate and the inherent response delay, thereby improving the accuracy.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description and the drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method for detecting coolant leakage in a liquid-cooled radiator according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a liquid coolant leakage detection system for a liquid-cooled radiator according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements.

[0020] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0021] Liquid-cooled radiators are devices that use circulating coolant to dissipate heat. Liquid-cooled systems utilize pumps to circulate coolant within the heat pipes for heat dissipation.

[0022] In related technologies, traditional methods for detecting coolant leaks in liquid-cooled radiators often result in incorrect leak locations when diagnosing minor leaks in tandem liquid-cooled radiator arrays. This is primarily because when micron-sized suspended particles are present in the coolant and are not removed by existing filtration mechanisms, these particles form a non-uniform, time-varying deposition layer on the pressure-sensitive diaphragm surfaces of various pressure measuring devices. This deposition layer then causes unique reading deviations and response delays in each device, below the hardware fault alarm threshold. The system's self-test program cannot recognize this complex device performance degradation, preventing it from accurately locating and diagnosing minor leaks in the piping.

[0023] For example, suppose a closed-loop coolant circulation system consisting of multiple sets of liquid-cooled radiators deployed in series is operating inside a large data center. A novel coolant additive is introduced, which polymerizes with the existing corrosion-resistant components in the coolant to generate micron-sized colloidal suspended particles. These particles cannot be effectively intercepted by existing filtration devices and selectively adhere to the pressure-sensitive diaphragms of multiple pressure measuring devices in the pipeline during coolant circulation. This adhesion causes a slight negative offset in the readings of each pressure measuring device and a millisecond-level delay in response to pressure changes within the pipeline. Due to the characteristics of the series deployment, the fluid environment of each device differs, resulting in varying particle deposition rates and thicknesses, leading to unique and unpredictable reading offsets and response delays for each pressure measuring device. When a connection in the pipeline develops a crack imperceptible to the naked eye due to minute vibration, causing coolant to leak at an extremely low rate, this real local pressure difference signal, after propagating and being received by these pressure measuring devices with different offsets and delays, ultimately results in a severely distorted pressure difference data matrix. When the differential pressure matrix analysis and location calculation method inside the system processed this distorted data, its calculation results deviated, incorrectly pointing the source of the leak to a non-real leak location.

[0024] If these issues are not addressed, the system will continue to face the risk of inaccurate leak location. Incorrect leak diagnosis will force maintenance teams to spend considerable time investigating non-leaked areas, resulting in significant service interruptions and wasted manpower. Simultaneously, real, minute leaks will persist undetected, potentially posing a long-term threat to the secure operation of the data center. The system will lose its ability to effectively perceive its true state, hindering timely corrective action and impacting equipment stability and the overall reliability of the data center.

[0025] Faced with the aforementioned problems, this application initially considered addressing particle deposition by improving the hardware accuracy of the pressure measurement device or modifying the filtration system. However, improving hardware accuracy typically comes at a high cost and is difficult to implement in retrofitting existing systems. Improving the filtration system may require downtime for modification, and there are technical bottlenecks in improving the filtration efficiency for micron-sized colloidal particles. Therefore, this application further considers whether, given the limitations of hardware-level modifications, a data processing and analysis approach could be adopted to compensate and correct the output data of the existing pressure measurement device, thereby overcoming the impact of performance degradation. This application considers dynamically analyzing pressure readings to capture leakage events and quantifying and compensating for the inherent response characteristics of the device to correct the timing information of leakage events. Finally, this information is combined to locate the leak point. This method can improve the accuracy of leak detection through software algorithm optimization without changing the existing hardware.

[0026] The embodiments of this application will be explained in detail below with reference to the accompanying drawings: Figure 1 This is an optional flowchart of the liquid coolant leakage detection method for liquid-cooled radiators provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S101 to S106.

[0027] Step S101: Obtain the pressure reading from the pressure measuring device; Step S102: Calculate the rate of pressure change using differential calculation based on the pressure reading; Step S103: Identify the start time of the leakage event based on the rate of pressure change; Step S104: Calculate the inherent response delay of the pressure measuring device; Step S105: Based on the inherent response delay, compensate for the occurrence sequence of the leakage event to obtain the compensated sequence. Step S106: Determine the location of the target leak point based on the starting time point and compensation sequence.

[0028] Steps S101 to S106 as shown in the embodiments of this application can combine the pressure change rate and the inherent response delay to achieve coolant leakage detection, thereby improving accuracy.

[0029] In some embodiments, steps S101-S106 aim to address the problem in the prior art where reading deviations and response delays in pressure measuring devices caused by suspended particles in the coolant lead to inaccurate leak location. By introducing a calculation of the inherent response delay of the pressure measuring device and compensating for the timing of leak events based on this calculation, measurement errors caused by equipment performance degradation are effectively eliminated, achieving the effect of accurately identifying the location of minute leak points under complex operating conditions.

[0030] First, pressure readings from the pressure measuring device are acquired. Based on these readings, differential calculations are used to calculate the rate of pressure change, transforming the static pressure value into dynamic trend information. This allows even minute pressure fluctuations to be effectively captured, enhancing sensitivity to potential leaks. Then, the onset time of the leak event is identified based on the rate of pressure change. This time point is the critical moment when the leak signal is first detected at a specific measuring device, providing an important benchmark for subsequent time analysis. Next, the inherent response delay of the pressure measuring device is calculated. This delay is the time lag in the device's response to pressure change signals, and its accurate quantification is crucial for eliminating measurement errors. Based on the inherent response delay, the timing of the leak event is compensated to obtain the compensated timing sequence. This compensation process corrects the time deviation caused by the device's response delay, ensuring that the leak event times recorded at different measuring points more accurately reflect the actual physical propagation process, guaranteeing the synchronization and accuracy of time information. Finally, the target leak point location is determined based on the onset time point and the compensated timing sequence. By comprehensively analyzing the pressure wave propagation information from multiple measuring points on the calibration time axis, the source of the pressure wave, i.e., the specific location of the leak point, can be inferred using the principles of sound wave propagation or time difference localization algorithms.

[0031] Understandably, inherent response delay refers to the time lag between receiving an actual pressure change signal and outputting a corresponding electrical signal from a pressure measuring device. This delay can be caused by factors such as the device's internal physical structure, the material properties of the sensing diaphragm, or the response speed of the signal processing circuit. Its purpose is to quantify and correct the inherent time deviation of the device's signal response, thereby improving the accuracy of the leak event timestamp. Compensated timing sequence, on the other hand, refers to the time sequence after correcting for the actual occurrence time of the leak event, taking into account the inherent response delay of the pressure measuring device. It can be obtained by subtracting or adding the corresponding inherent response delay to the original timestamp. Its purpose is to eliminate the impact of device response delay on the judgment of leak event time, ensure time synchronization between different measurement points, and provide calibrated time information for accurate leak location.

[0032] To illustrate this technical solution more clearly, a specific example is provided below. First, the pressure measuring device can employ a high-precision piezoresistive pressure sensor, which converts the pressure signal in the pipeline into a digital pressure reading via an analog-to-digital converter and transmits it to the central processing unit at a sampling frequency of 1000 times per second. In the central processing unit, the pressure reading is received in real time. To calculate the rate of pressure change, a moving average difference algorithm can be applied to a continuous sequence of pressure readings. For example, the difference between the current reading and the average of the previous 10 sampling points can be calculated and divided by the corresponding time interval. When the absolute value of the rate of pressure change exceeds a preset dynamic threshold, the system triggers a leakage signal and performs local waveform analysis on the signal. For example, within a preset 50-millisecond time window, the first significant inflection point or local extreme point in the waveform is identified, and its corresponding timestamp is used as the start time of the leakage event. To calculate the inherent response delay of the pressure measuring device, a reference pressure pulse of known waveform can be injected into the pipeline periodically or at system startup, while simultaneously acquiring the reference pressure waveform data segment and the pressure waveform data stream to be matched. By performing cross-correlation on these two data segments, the time offset corresponding to the peak value of the largest amplitude in the cross-correlation function is the actual propagation time difference. Combining the physical distance between adjacent pressure measuring devices and the theoretical speed of sound of the coolant, the theoretical propagation time difference can be calculated. The difference between the two is the inherent response delay of the device. After obtaining the inherent response delay of each pressure measuring device, the system subtracts its corresponding inherent response delay from the start time of the leak event detected by each device, thus obtaining the compensated leak event occurrence sequence. For example, if a device detects a leak at a start time of T1 and its inherent response delay is D1, then the compensated sequence is T1-D1. Finally, using these compensated start times, combined with the geometry of the cooling pipes and the speed of sound of the coolant, the Time Difference of Origin (TDOA) algorithm can be used to determine the location of the target leak point. For example, by comparing the differences in the compensated sequence of leak events detected by different pressure measuring devices, and combining this with the propagation path and speed of the pressure wave in the pipes, the common source of the pressure wave, i.e., the leak point, can be deduced.

[0033] Through the above technical solution, this embodiment effectively solves the problem of inaccurate leak location caused by the performance degradation (reading offset and response delay) of pressure measurement equipment due to micron-sized suspended particles in the coolant, which is present in the prior art. By sensitively capturing the rate of pressure change, combined with accurate calculation and timing compensation of the inherent response delay of the equipment, this embodiment can correct the time deviation of the measurement data, thereby ensuring the authenticity of the timing of the leak event. This enables the system to locate leaks based on more accurate time information, avoiding the situation of mistakenly pointing the leak source to other locations, significantly improving the accuracy of leak diagnosis, reducing unnecessary troubleshooting time and maintenance costs, and ensuring the stable operation of the liquid cooling system.

[0034] In some embodiments, step S102, calculating the pressure change rate using differential calculation based on the pressure reading, may include, but is not limited to, the following steps: Obtain historical stress data; Characteristic analysis is performed on historical pressure data and pressure readings to obtain fluctuation characteristics, which include fluctuation amplitude or frequency distribution. Determine the time interval for difference operations based on the fluctuation characteristics; The rate of pressure change is calculated based on historical pressure data, pressure readings, and differential calculation time intervals.

[0035] In some embodiments, pressure readings may fluctuate due to various factors in actual liquid cooling systems (such as coolant disturbances and the inherent accuracy limitations of the pressure sensor). If differential calculations are performed directly using the pressure readings, these fluctuations may be amplified, leading to inaccurate calculated pressure change rates and consequently affecting the identification of the leak's initiation time and the location of the leak. Therefore, it is necessary to eliminate the impact of pressure reading fluctuations on differential calculations and improve the accuracy of pressure change rate calculations.

[0036] First, historical pressure data can be acquired. Feature analysis of the historical pressure data and pressure readings yields fluctuation characteristics, which quantifies the noise or disturbance characteristics present in the current pressure signal. For example, a large fluctuation amplitude indicates strong pressure signal noise; if the frequency distribution shows high-frequency components, rapid disturbances may exist. Fluctuation characteristics include fluctuation amplitude or frequency distribution. Based on these fluctuation characteristics, the differential calculation time interval is determined. When large pressure fluctuations or high-frequency noise are detected, a larger time interval can be selected for differential calculation, effectively smoothing out these instantaneous fluctuations and noise, and preventing them from interfering with the calculation of the pressure change rate. Conversely, when pressure fluctuations are small or noise levels are low, a smaller time interval can be used to more precisely capture pressure changes caused by real events such as leaks. The pressure change rate is calculated based on historical pressure data, pressure readings, and the differential calculation time interval. This adaptive adjustment of the differential calculation time interval ensures that the calculated pressure change rate effectively suppresses noise, thus more accurately reflecting the true pressure change trend within the pipeline.

[0037] Understandably, fluctuation amplitude refers to the difference between the maximum and minimum values ​​of pressure data within a certain time range, or the degree to which pressure data deviates from its average value. This can be obtained by calculating the standard deviation, root mean square value, or peak-to-peak value, and its purpose is to quantify the severity of pressure fluctuations. Frequency distribution refers to the energy or intensity distribution of different frequency components in the pressure data. This can be obtained through signal processing methods such as Fourier transform and wavelet analysis, and its purpose is to reveal the periodicity of the main noise components in pressure fluctuations. The differential operation time interval refers to the time span between sampling two pressure readings when calculating the rate of pressure change. This can be set based on the sampling frequency of the pressure data, the system response time, or a preset empirical value, and its purpose is to influence the sensitivity of differential operations to noise and its ability to capture real pressure changes.

[0038] To illustrate this technical solution more clearly, a specific example is provided below. In a liquid cooling system, a pressure measurement device can acquire pressure readings at a frequency of 100 times per second and store these readings in a circular buffer to form historical pressure data. When it is necessary to calculate the current pressure change rate, the system can retrieve the most recent 1000 pressure readings from this buffer as historical pressure data. To perform feature analysis on the historical pressure data and the current pressure readings, the following method can be used: First, a Fast Fourier Transform (FFT) is performed on the historical pressure data to obtain its frequency distribution and identify the presence of high-frequency noise components. Simultaneously, the standard deviation of the historical pressure data within a sliding window can be calculated to quantify the fluctuation amplitude. For example, if the standard deviation exceeds a preset threshold, or the FFT result shows significant energy in a specific high-frequency band, it can be determined that the current pressure signal exhibits large fluctuation characteristics. Based on these fluctuation characteristics, the differential operation time interval can be dynamically determined. For example, if high-frequency noise or large fluctuation amplitude is detected, the system can increase the differential operation time interval, for example, from the default 0.1 seconds to 0.5 seconds. This means that when calculating the pressure change rate, the current pressure reading will be compared with the pressure reading 0.5 seconds ago. If the fluctuations are small, a smaller time interval, such as 0.1 seconds, can be maintained to capture finer pressure changes. Finally, using a determined differential calculation time interval, combined with historical pressure data and the current pressure reading, the pressure change rate is calculated. For example, if the current pressure reading is [value] and the time interval is Δt, the pressure reading corresponding to the time point (-Δt) can be found from the historical data, and then the pressure change rate can be calculated as (-) / Δt, where [value] represents the current time. This adaptive adjustment ensures high accuracy in the calculated pressure change rate under different noise conditions.

[0039] Through the above technical solution, in liquid cooling systems, the influence of pressure reading fluctuations on differential calculations can be effectively eliminated when calculating the rate of pressure change based on pressure readings. By performing feature analysis on historical pressure data and pressure readings, and adaptively determining the differential calculation time interval based on fluctuation characteristics, noise interference with the calculation results can be reduced, thereby improving the accuracy of pressure change rate calculation. This accurate calculation of the rate of pressure change provides a reliable data foundation for subsequently identifying the start time of leak events and locating leak points, avoiding misjudgments caused by data fluctuations, and improving the overall performance of the leak detection system.

[0040] In some embodiments, in step S103, identifying the start time of the leakage event based on the rate of pressure change may include, but is not limited to, the following steps: If the rate of pressure change is greater than a preset threshold, a leakage signal is acquired. Local waveform analysis is performed on the leakage signal to obtain the local waveform of the signal; Within a preset time window, identify target points in the local waveform of the signal. Target points include waveform inflection points or local extreme points. Use the timestamp corresponding to the target point as the starting time point.

[0041] In some embodiments, since the rate of pressure change may be affected by various factors, such as normal pressure fluctuations within the system and sensor noise, simply using a pressure change rate exceeding a threshold as the criterion for determining a leakage signal can easily lead to misjudgment. Furthermore, leakage signals may suffer from waveform distortion and noise interference, all of which affect the accuracy of the start time point identification. Therefore, it is necessary to accurately determine the start time point of the leakage event. This can be achieved by first determining whether the rate of pressure change exceeds a preset threshold. If the rate of pressure change exceeds the preset threshold, a preliminary judgment is made that a leakage may exist, and a segment of the leakage signal is obtained. This effectively filters out interference caused by normal system fluctuations or sensor noise, focusing the analysis on the potential leakage event. Then, local waveform analysis is performed on the leakage signal to obtain the local waveform, allowing for a deeper understanding of the inherent characteristics of the leakage signal, such as its shape, trend, and dynamic changes, rather than simply judging the amplitude. Within a preset time window, target points in the local waveform are identified, including waveform inflection points or local extreme points. Waveform inflection points typically indicate significant shifts in the rate of pressure change, while local extrema may represent peaks or troughs in the effects of a leak. These points are physically closely related to the actual occurrence of the leak event or its critical phases. The timestamp corresponding to the target point is then used as the starting time point. By identifying these points with specific physical significance within a defined time window, it is possible to avoid misinterpreting subsequent signal characteristics unrelated to the initial point as the starting point.

[0042] Understandably, a leak signal refers to a sequence of raw or preliminarily processed pressure data acquired from a pressure measuring device when the rate of pressure change exceeds a preset threshold, indicating a potential leak event. Specifically, it can be a sequence of pressure readings over a period of time, or pressure data that has undergone preliminary filtering. Its purpose is to provide a data foundation for subsequent refined analysis, ensuring that the analysis focuses on the time period in which a leak may occur. A local waveform refers to waveform data extracted from the leak signal through local waveform analysis, reflecting the pressure change characteristics during a specific time period. Specifically, it can be a pressure time series segment that has undergone smoothing, denoising, or feature enhancement processing. Its purpose is to provide a clear waveform representation for further analysis, facilitating the identification of key target points.

[0043] To illustrate this technical solution more clearly, a specific example is used below. After continuously acquiring pressure readings from the pressure measuring device and calculating the rate of pressure change, the processor compares this rate in real time with a preset threshold. For example, this preset threshold can be set to three standard deviations of the normal pressure fluctuation range, based on statistical analysis of historical pressure data under leak-free conditions. Once the rate of pressure change consistently exceeds this threshold, a program to acquire a leak signal is triggered. This leak signal can be a sequence of pressure readings including 5 seconds before and after the threshold trigger point. Subsequently, the processor performs local waveform analysis on the acquired leak signal. Specifically, wavelet transform can be used to decompose the leak signal into multiple scales to remove high-frequency noise and highlight the transient characteristics of the signal, thereby obtaining a local waveform. After obtaining the local waveform, target points can be identified within a preset time window. This preset time window can be set to within 10 seconds from the first time the rate of pressure change exceeds the threshold. When identifying target points, a second derivative-based method can be used to detect waveform inflection points, specifically finding points where the sign of the second derivative changes and the first derivative is not zero. Simultaneously, a sliding window maximum / minimum detection algorithm can be used to identify local extrema. For example, when an inflection point is detected where the waveform changes from a downward trend to an upward trend, or when the pressure value reaches a local minimum, these points are marked as target points. Finally, the timestamp corresponding to the first identified target point that meets leakage characteristics (e.g., the starting point of a continuous pressure decrease) is recorded and output as the starting time point of the leakage event. For example, if multiple target points are detected, the system can prioritize selecting the inflection point with the most significant pressure decrease trend as the starting point, or the starting point where the pressure value first significantly decreases among the local extrema.

[0044] Through the above technical solution, this embodiment effectively solves the problem of misjudgment that easily occurs when relying solely on pressure change rate thresholds in existing technologies, and overcomes the impact of leakage signal waveform distortion and noise interference on the accuracy of starting time point identification. This embodiment introduces local waveform analysis of the leakage signal and combines this with identifying waveform inflection points or local extreme points within a preset time window as target points, enabling precise capture of the true start time of the leakage event from complex pressure data. This allows the system to more accurately determine the occurrence of a leak, significantly improving the accuracy of leak event starting time point identification, thereby providing more reliable and refined time information for subsequent leak location and fault diagnosis, and avoiding misjudgments and resource waste caused by inaccurate starting time points.

[0045] In some embodiments, calculating the inherent response delay of the pressure measuring device in step S104 may include, but is not limited to, the following steps: Acquire reference pressure waveform data segments and pressure waveform data streams to be matched; The cross-correlation function is obtained by performing cross-correlation calculations on the reference pressure waveform data segment and the pressure waveform data stream to be matched. The actual propagation time difference is determined by the time offset corresponding to the peak point with the largest amplitude in the cross-correlation function; Calculate the theoretical propagation time difference between adjacent pressure measuring devices based on the physical distance between them and the sound velocity of the coolant. Calculate the inherent response delay based on the actual propagation time difference and the theoretical propagation time difference.

[0046] In some embodiments, because the liquid-cooled heat sink assembly is deployed in series, micron-sized suspended particles in the coolant may form an uneven, time-varying deposition layer on the pressure-sensitive diaphragm surface of the pressure measuring device. This causes each pressure measuring device to generate its own reading offset and response delay below the hardware fault alarm threshold. In this case, simply correcting the timing using the theoretically calculated propagation time difference cannot accurately reflect the actual response delay, thus affecting the accuracy of leak location.

[0047] To address this, a reference pressure waveform data segment and a pressure waveform data stream to be matched are first acquired. Cross-correlation is then performed on these two data streams to obtain a cross-correlation function, which quantifies the similarity between the two signals at different time offsets. Then, based on the time offset corresponding to the peak point with the largest amplitude in the cross-correlation function, the actual propagation time difference is determined. This reflects the propagation characteristics of pressure waves in a real pipeline environment, incorporating the influence of various practical factors such as medium characteristics, pipeline structure, and equipment response. Next, based on the physical distance between adjacent pressure measuring devices and the sound velocity of the coolant, the theoretical propagation time difference between adjacent pressure measuring devices is calculated, providing a benchmark unaffected by the device's own response. Finally, the inherent response delay is calculated based on the actual propagation time difference and the theoretical propagation time difference. This embodiment overcomes the problem in existing technologies where the deposition of suspended particles in the coolant causes uneven, time-varying reading offsets and response delays in pressure measuring devices.

[0048] Understandably, a reference pressure waveform data segment refers to a pre-defined pressure waveform sample with specific characteristics, which serves as a benchmark signal for comparison with the actually acquired pressure waveform. Its purpose is to provide a known time or waveform feature point for subsequent time delay calculation. The pressure waveform data stream to be matched refers to a continuous sequence of pressure data acquired in real-time or near real-time from the pressure measuring device, containing pressure fluctuation information that may be caused by leakage events. Its purpose is to provide the actual pressure change signal for comparison with the reference waveform to reveal temporal differences. The time offset refers to the amount by which two signals are offset from each other on the time axis. When the cross-correlation function reaches its maximum value, this offset is the time difference between the two signals that best match. It can be measured in milliseconds or microseconds. Its purpose is to quantify the time difference between the arrival times of two pressure waveforms at different measuring devices.

[0049] To illustrate this technical solution more clearly, a specific example is used below. First, a known, instantaneous pressure pulse can be used as a reference pressure waveform data segment. This pulse can be generated by the system in a specific maintenance mode through rapid valve opening and closing or starting of a small pump, and acquired by a calibrated standard pressure sensor. Simultaneously, the output pressure signal is continuously acquired from the pressure measuring device to be calibrated (e.g., an adjacent device downstream of the reference sensor), forming a pressure waveform data stream to be matched. Next, the acquired reference pressure waveform data segment and the pressure waveform data stream to be matched are input into a signal processing unit, which can be an embedded processor or an industrial computer. This processing unit performs cross-correlation operations to generate a cross-correlation function. For example, a Fast Fourier Transform (FFT) can be used for frequency domain cross-correlation to improve computational efficiency. The result of the cross-correlation function will be displayed as a time series containing one or more peaks. Subsequently, the signal processing unit identifies the peak point with the largest amplitude in the cross-correlation function and extracts the time offset corresponding to that peak point. This time offset is the actual time it takes for the pressure wave to propagate from the reference sensor to the sensor to be calibrated, i.e., the actual propagation time difference. Simultaneously, the system pre-stores the physical distances between adjacent pressure measuring devices, for example, obtained through CAD drawings or on-site measurements. The sound velocity of the coolant can be obtained by looking up tables or calculating in real time based on the coolant's type, temperature, and pressure. The signal processing unit uses these parameters to calculate the theoretical propagation time difference of the pressure wave from one device to another under ideal conditions through simple division. Finally, the signal processing unit compares the actual propagation time difference with the theoretical propagation time difference; the difference is the inherent response delay of the pressure measuring device to be calibrated. This delay value can be stored in the device's calibration parameter database for subsequent leakage event timing compensation. In this way, even if the pressure measuring device's response characteristics change due to long-term operation or media contamination, its current true inherent response delay can be obtained through periodic or on-demand calibration.

[0050] Through the above technical solution, this embodiment can accurately calculate the inherent response delay of pressure measuring devices. By performing cross-correlation analysis on the actual collected pressure waveform data, the propagation time of pressure waves in the real pipeline environment can be accurately captured and compared with the theoretical propagation time, thereby quantifying the response lag of each pressure measuring device caused by factors such as deposits. This allows the system to obtain more accurate device response characteristic parameters, providing a reliable basis for subsequent leakage event timing compensation, thus significantly improving the accuracy of leak location and avoiding misjudgments and resource waste caused by device response delays.

[0051] In some embodiments, in step S106, determining the location of the target leak point based on the start time point and compensation sequence may include, but is not limited to, the following steps: Step S201: Obtain the physical properties of the coolant in the cooling pipes; Step S202: Calculate the sound velocity of the coolant in the cooling pipes based on the physical properties of the coolant. Step S203: Set multiple candidate leak point locations; Step S204: Calculate the theoretical propagation time of the pressure wave from the candidate leak point to the pressure measuring device based on the sound velocity of the coolant; Step S205: Identify the location of the first leak point based on the starting time and the rate of pressure change; Step S206: Identify the location of the second leak point based on the theoretical propagation time and compensation sequence; Step S207: Determine the location of the target leak point based on the location of the first leak point and the location of the second leak point.

[0052] In some embodiments, due to the complexity of cooling pipelines and the variability of coolant physical properties, relying solely on the start time point and compensation timing is insufficient to accurately locate leak points, potentially leading to discrepancies between the location results and the actual leak location. To address this, the physical properties of the coolant in the cooling pipelines can be obtained first, and the coolant sound velocity can be calculated based on these properties. Then, multiple candidate leak point locations are set, and for each candidate point, the theoretical propagation time of the pressure wave from the candidate leak point location to the pressure measuring device is calculated based on the coolant sound velocity, thus providing necessary data support for accurately simulating the behavior of pressure waves in the pipeline. Based on this, this embodiment employs two different strategies in parallel to identify leak points. On one hand, the first leak point location is identified based on the start time point and pressure change rate—a preliminary judgment based on the initial characteristics of a leak event. On the other hand, the second leak point location is identified based on the theoretical propagation time and compensation timing—a precise location based on a pressure wave propagation physical model and accurate time correction. Finally, the target leak point location is determined based on the first and second leak point locations. This dual identification and fusion determination strategy significantly enhances the accuracy and robustness of the location.

[0053] Understandably, the physical properties of the coolant in cooling pipes refer to the inherent attributes of the coolant within the pipes, such as density, viscosity, compressibility, temperature, and pressure. These properties can be obtained through direct measurement (e.g., real-time data acquisition via sensors), lookup tables (e.g., consulting a pre-defined database based on coolant type and operating conditions), or calculation based on fluid dynamics models. The purpose is to provide an accurate physical parameter basis for subsequent sound velocity calculations. The coolant sound velocity in cooling pipes refers to the speed at which pressure waves propagate within the coolant. This can be obtained through calculations based on the coolant's bulk modulus and density, experimental measurements, or by consulting coolant sound velocity characteristic curves. The purpose is to provide key parameters for calculating the propagation time of pressure waves.

[0054] To illustrate this technical solution more clearly, a specific example is provided below. Obtaining the physical properties of the coolant in the cooling pipeline can be achieved by deploying temperature, pressure, and flow sensors at key nodes in the cooling pipeline to collect real-time data on the coolant's temperature, pressure, and flow rate. This data can be combined with a pre-established table of coolant thermodynamic properties and interpolation or fitting algorithms to dynamically calculate the coolant's density, viscosity, bulk modulus, and other physical properties under current operating conditions. Based on the coolant's physical properties, the coolant's sound velocity in the cooling pipeline can be calculated using the obtained coolant density and bulk modulus, employing the sound velocity formula (e.g., sound velocity equals the ratio of the bulk modulus to the square root of the density). Setting multiple candidate leak point locations can be done by pre-marking all pipe connections, elbows, valves, and radiator interfaces—areas prone to leakage—as candidate leak point locations in the CAD model or topology diagram of the cooling pipeline. These locations can be assigned unique identifiers and corresponding coordinate information.

[0055] Then, based on the coolant sound velocity, the theoretical propagation time of the pressure wave from the candidate leak point to the pressure measuring device is calculated. Specifically, for each candidate leak point, the actual distance of the pipeline from it to each pressure measuring device is used, combined with the calculated coolant sound velocity, and the theoretical propagation time of the pressure wave from that candidate point to each pressure measuring device is calculated using a simple formula of distance divided by sound velocity. Based on the starting time point and the rate of pressure change, the location of the first leak point can be identified by analyzing the spatial distribution of the rate of pressure change recorded by each pressure measuring device near the starting time point of the leak event. For example, the rate of pressure change near the leak point usually exhibits a larger absolute value or a steeper decreasing trend. By performing spatial interpolation or gradient analysis on these rate data, the area of ​​most drastic pressure change can be preliminarily located, thereby identifying the location of the first leak point.

[0056] Next, based on the theoretical propagation time and compensation timing, the location of the second leak point is identified. This can be done by comparing the actual arrival time of the pressure wave received by each pressure measuring device (after compensation timing correction) with the theoretical propagation time from each candidate leak point location to these devices. By minimizing the error between the actual arrival time and the theoretical propagation time, for example, using the least squares method or triangulation algorithm, the candidate point that best conforms to the pressure wave propagation law can be identified as the location of the second leak point. Based on the locations of the first and second leak points, the target leak point location can be determined by weighted averaging or fusion of the identified first and second leak point locations. For example, different weights can be assigned based on the confidence level or historical accuracy of the two location methods, and then the final target leak point location can be calculated. Alternatively, if the two locations are close together, their geometric center can be taken; if they are far apart, further verification or selection of the one with higher confidence may be necessary.

[0057] Through the above technical solution, this embodiment, based on determining the target leak point location according to the starting time point and compensation timing sequence, further introduces the acquisition of the physical properties of the coolant in the cooling pipe and the calculation of the coolant sound velocity, making the pressure wave propagation model closer to reality. By setting multiple candidate leak point locations and calculating the theoretical propagation time, a reference benchmark is provided for accurate positioning. Simultaneously, the first leak point location is identified by combining the starting time point and pressure change rate, and the second leak point location is identified using the theoretical propagation time and compensation timing sequence, achieving multi-dimensional, multi-path preliminary leak point identification. Finally, by comprehensively considering these two identification results, the positioning accuracy and reliability of the target leak point can be significantly improved, effectively avoiding positioning deviations caused by a single information source or simplified model, thereby reducing false alarms and unnecessary troubleshooting work, and ensuring the stable operation of the liquid cooling system.

[0058] In some embodiments, in step S205, identifying the location of the first leak point based on the starting time point and the rate of pressure change may include, but is not limited to, the following steps: The absolute value of the rate of pressure change is taken as the event intensity; Spatiotemporal correlation analysis was performed on multiple starting time points to obtain multiple pressure wave propagation paths; Identify the changing trends of event intensity in each pressure wave propagation path; If the trend is downward, the location of the first leak point can be identified based on the starting time and the laws of fluid physics propagation.

[0059] In some embodiments, the rate of pressure change can be affected by various factors, such as the flow state of the coolant, pipeline structure, and environmental noise. These factors can cause fluctuations or abrupt changes in the rate of pressure change, thus affecting the accuracy of leak location identification. Furthermore, since the propagation of pressure waves in the coolant is constrained by the laws of fluid physics, locating leaks based on the starting time point may ignore the influence of these physical laws, leading to inaccurate location results. Therefore, it is necessary to utilize the starting time point and the rate of pressure change, combined with the laws of fluid physics propagation, to improve the accuracy of leak location identification.

[0060] The absolute value of the pressure change rate can be used as the event intensity to quantify the extent of the leak and provide a quantitative standard for subsequent analysis. Then, spatiotemporal correlation analysis is performed on multiple starting time points to obtain multiple pressure wave propagation paths. This considers not only temporal information but also spatial location information, thus understanding the propagation range and impact of the pressure waves. Next, the trend of event intensity change in each pressure wave propagation path is identified. Since pressure waves gradually attenuate due to energy dissipation during propagation in fluids, a decreasing trend in event intensity provides evidence that the leak source is located at the starting point of that path. If the trend is decreasing, the location of the first leak point can be identified based on the starting time point and fluid physics propagation laws, such as the propagation speed and attenuation characteristics of pressure waves in coolant. This overcomes the limitations of traditional methods that rely solely on a single time point or pressure threshold for judgment.

[0061] Understandably, event intensity refers to an indicator used to quantify the severity of a leak event. It can be characterized by the absolute value of the rate of pressure change, reflecting the extent of the leak. Pressure wave propagation path refers to the actual or inferred path taken by the pressure wave from the leak source to various pressure measuring devices. It can be constructed using pipeline topology combined with the pressure wave arrival sequence, aiming to depict the diffusion range of the pressure wave in the system. The trend of event intensity refers to the variation of the pressure wave intensity with distance or time during propagation. It can be identified using linear regression, curve fitting, or difference calculation, aiming to determine the attenuation characteristics of the pressure wave. Fluid physics propagation laws refer to the physical laws and models describing the propagation behavior of pressure waves in a fluid medium. They can be represented by models of sound wave propagation speed in fluids, energy attenuation models, or fluid dynamics wave equations, aiming to provide a physical basis for inferring the leak point.

[0062] To illustrate this technical solution more clearly, a specific example is used below. First, when a pressure change is detected, the rate of pressure change recorded by each pressure measuring device within a specific time window is calculated, and its absolute value is taken as the event intensity at that time point. For example, if the rate of pressure change of a certain pressure measuring device at a certain moment is -0.5 MPa / s, then its event intensity is set to 0.5 MPa / s.

[0063] Next, to perform spatiotemporal correlation analysis on multiple starting time points, a pipeline topology graph can be constructed, where nodes represent pressure measuring devices and edges represent connecting pipelines. Once multiple pressure measuring devices detect leakage signals and record their respective starting time points, a graph-based shortest path algorithm or time synchronization algorithm, combined with the physical distances between devices and the approximate speed of sound of the coolant, can be used to infer the theoretical time difference of pressure wave propagation from a possible leak source to each device. By comparing the degree of matching between the actual starting time points and the theoretical time differences, multiple possible pressure wave propagation paths can be identified. For example, if devices A, B, and C detect signals sequentially, and the time differences match their order and distance in the pipeline, a propagation path from A to B and then to C can be inferred.

[0064] Subsequently, for each identified pressure wave propagation path, the event intensity recorded by each pressure measuring device along the path can be analyzed. For example, the event intensity of adjacent devices along the path can be compared, or a linear fit can be performed on the event intensity data for the entire path. If the fitting result shows that the event intensity decreases with increasing propagation distance, for example, the event intensity value continuously decreases from the beginning to the end of the path, then the path is considered to conform to the characteristics of leakage wave attenuation.

[0065] Finally, if the event intensity change trend along a pressure wave propagation path is confirmed to be decreasing, the location of the first leak point can be identified based on the starting time of the path and a pre-established fluid physics propagation law model. For example, using the pressure wave attenuation model and propagation velocity model in the coolant, combined with the location and signal strength of the first pressure measuring device that detects the signal along the path, as well as the signal strength and arrival time of subsequent devices, iterative calculations or optimization algorithms can be used to determine the location of leak points that conform to these propagation characteristics.

[0066] Through the above technical solution, this embodiment can convert the rate of pressure change into quantified event intensity, and by combining spatiotemporal correlation analysis of multiple starting time points, multiple pressure wave propagation paths can be obtained. By identifying the changing trends of event intensity in these paths and filtering out paths with a decreasing trend, pressure fluctuations caused by noise or non-leakage factors can be eliminated. Finally, based on the starting time point and the laws of fluid physics propagation, the accuracy of leak point location can be improved, avoiding misjudgments caused by pressure data fluctuations or neglect of physical laws, and providing preliminary location information for subsequent leak point determination.

[0067] In some embodiments, in step S201, obtaining the physical properties of the coolant in the cooling pipes may include, but is not limited to, the following steps: Step S301: Divide the cooling pipes into multiple pipe sections; Step S302: Collect the physical property data corresponding to each pipe segment; Step S303: Calculate the physical properties of the coolant in a single pipe section based on the fluid path, heat exchange characteristics, and physical property data of the pipe section; Step S304: Perform time smoothing on the physical properties of the coolant in multiple single pipe sections to obtain the physical properties of the coolant in the cooling pipe.

[0068] In some embodiments, because actual liquid cooling systems often have complex and long cooling pipes, the physical properties of the coolant may differ in different pipe sections. Simply treating the entire cooling pipe as a whole to obtain the coolant's physical properties ignores the differences in coolant physical properties between different pipe sections, which reduces the accuracy of leak location. Therefore, the cooling pipe can be divided into multiple pipe sections. This is because the physical properties of the coolant may differ significantly in different pipe sections due to differences in fluid path and heat exchange characteristics. By segmenting, the distribution of coolant physical properties can be described more precisely. Then, physical property data corresponding to each pipe section is collected. This data is the basis for calculating the coolant physical properties of a single pipe section. Different pipe sections may require different physical property data, such as temperature, pressure, and flow rate, to comprehensively reflect the state of the coolant within that section. Then, based on the fluid path, heat exchange characteristics, and physical property data of the pipe section, the physical properties of the coolant in a single pipe section are calculated. The fluid path determines the flow pattern of the coolant within the pipe section, the heat exchange characteristics determine the heat exchange between the coolant and the surrounding environment, and the physical property data reflect the actual state of the coolant. Considering these three factors comprehensively allows for a more accurate calculation of the physical properties of the coolant in a single pipe section. Finally, time smoothing is applied to the physical properties of the coolant in multiple single pipe sections to obtain the physical properties of the coolant in the entire cooling system. Time smoothing eliminates fluctuations caused by measurement errors or instantaneous disturbances, making the obtained physical properties of the coolant in the cooling system more stable and reliable. By comprehensively considering the physical properties of multiple single pipe sections, the overall physical properties of the entire cooling system can be obtained.

[0069] As can be understood, fluid path refers to the flow trajectory and mode of coolant within a pipe section. Specifically, this includes the flow direction, velocity distribution, the presence of turbulent or laminar flow regions, and the influence of pipe geometry on fluid flow. Its purpose is to reflect the dynamic behavior of the coolant within a specific pipe section. Heat exchange characteristics refer to the properties of heat transfer between the coolant and the external environment or equipment within the pipe section. Specifically, this includes the thermal conductivity of the pipe material, pipe wall thickness, surface area, the ambient temperature of the pipe section, and the presence of heat sinks, among other factors. Its purpose is to reflect the temperature change pattern of the coolant within a specific pipe section.

[0070] To illustrate this technical solution more clearly, a specific example is provided below. Cooling piping can be divided based on the connection points of key components (e.g., radiators, pumps, valves) or the installation locations of physical sensors (e.g., temperature sensors, pressure sensors), thus logically dividing the complex cooling piping system into several independent segments. For example, the piping from the pump outlet to the first radiator inlet can be defined as one segment, the flow channels within each radiator as another segment, and the connecting piping between radiators as yet another segment. Within each segment, corresponding sensors can be deployed to collect physical property data. For example, high-precision temperature and pressure sensors can be installed at the inlet and outlet of each segment to obtain real-time temperature and pressure data of the coolant. For segments requiring flow rate information, miniature flow meters can be installed. These sensors can collect data periodically or when triggered by specific events and transmit the data to the central processing unit.

[0071] When calculating the physical properties of coolant in a single pipe section, fluid dynamics and heat transfer models can be used. For example, for a specific pipe section, based on its geometry, material properties, initial temperature and pressure of the coolant, as well as the fluid path (e.g., straight pipe, elbow, reducer) and heat exchange characteristics (e.g., whether it is in contact with heat-generating equipment, whether there is external heat dissipation), combined with collected temperature and pressure data, the average density, viscosity, specific heat capacity, and other physical properties of the coolant within that pipe section can be calculated by solving the energy conservation equation and the mass conservation equation. For example, finite element analysis or computational fluid dynamics (CFD) simulation software can be used to simulate the flow and heat transfer process of the coolant within the pipe section, thereby obtaining a more refined distribution of physical properties and calculating the representative physical properties of that pipe section. When performing time-smoothing of the physical properties of coolant in multiple single pipe sections, digital signal processing techniques can be used. For example, a moving average filter can be applied to the sequence of coolant physical properties calculated for each pipe section over a period of time to eliminate instantaneous measurement noise or calculation errors. Alternatively, exponentially weighted moving average (EWMA) or Kalman filtering algorithms can be used. These algorithms can dynamically adjust the estimated values ​​of physical properties based on the historical trends of the data and the current measurements, thereby obtaining a smooth and stable physical property data of the coolant in the cooling pipes. This data can more accurately reflect the overall physical state of the entire cooling pipes at a specific point in time.

[0072] Through the above technical solution, this embodiment overcomes the limitations of simply treating the entire cooling pipeline as a whole to obtain the physical properties of the coolant. By segmenting the cooling pipeline and combining the fluid path, heat exchange characteristics, and actual collected physical property data of each segment for calculation, the true physical state of the coolant in different segments can be reflected more precisely and accurately. Furthermore, by performing time smoothing on these segmented calculation results, measurement noise and instantaneous fluctuations can be effectively eliminated, ensuring that the obtained coolant physical property data of the cooling pipeline is stable and reliable. This improves the accuracy of subsequent calculations of coolant sound velocity and determination of target leak point locations based on coolant physical properties, thereby effectively avoiding misjudgments of leak points due to inaccurate estimation of coolant physical properties and improving the reliability of leak detection in liquid cooling systems.

[0073] In some embodiments, in step S303, calculating the physical properties of the coolant in a single pipe section based on the fluid path, heat exchange characteristics, and physical property data of the pipe section may include, but is not limited to, the following steps: Step S401: Filter the physical property data; Step S402: Based on the filtered physical property data and the energy and mass conservation rules of the pipe section, verify the target parameters, including parameters of the fluid path or parameters of heat exchange characteristics. Step S403: Calculate the physical properties of the coolant in a single pipe section based on the verified target parameters.

[0074] In some embodiments, the actual collected physical property data may contain noise or errors, and directly using this data for calculations can lead to insufficient accuracy in the calculation results of the physical properties of the coolant in a single pipe section, thus affecting the accuracy of leak detection. Furthermore, uncertainties in the fluid path parameters or heat exchange characteristic parameters of the pipe section can also cause deviations in the calculation results. Therefore, the physical property data can be filtered first, effectively removing sensor noise and environmental interference, ensuring the purity and reliability of the data foundation upon which subsequent calculations rely. This allows subsequent parameter verification and physical property calculations to be built on a solid foundation, avoiding calculation deviations caused by errors in the original data. Then, based on the filtered physical property data and the energy and mass conservation rules of the pipe section, the target parameters are verified. This utilizes the fundamental physical laws of coolant flow within the pipe section, dynamically correcting the fluid path parameters or heat exchange characteristic parameters using actual measurement data. For example, if the calculated energy or mass is not conserved, the relevant parameters can be adjusted in reverse until the conservation conditions are met, thereby overcoming the inherent uncertainties or time-varying deviations of these parameters, making the model parameters used to calculate the physical properties of the coolant in a single pipe section closer to actual operating conditions. The target parameters include parameters of the fluid path or parameters of heat exchange characteristics. Based on the validated target parameters, the physical properties of the coolant in a single pipe section are then calculated. Because the input data has been filtered and the validation accuracy of the model parameters has been improved, the accuracy and reliability of the calculated physical properties of the coolant in a single pipe section are enhanced.

[0075] Understandably, physical property data refers to parameters reflecting the state of the coolant collected at various sections of the cooling pipeline, such as temperature, pressure, and flow rate. These can be acquired in real time using sensors or through other measuring devices. The energy and mass conservation rule for a pipe section refers to the physical laws governing the energy and mass of the coolant during input, output, and internal changes within a specific pipe section. That is, energy cannot be created or destroyed, and the total mass remains constant. This can be characterized using a mathematical model based on the first law of thermodynamics and the law of conservation of mass.

[0076] To illustrate this technical solution more clearly, a specific example is provided below. When filtering physical property data, a digital signal processor or microcontroller can be used to execute a Kalman filter algorithm to process the raw data obtained from temperature, pressure, and flow sensors in real time. For example, for pressure data, a dynamic Kalman gain can be set to adapt to noise levels under different operating conditions, effectively smoothing fluctuations while maintaining data responsiveness. When verifying target parameters based on the filtered physical property data and the energy and mass conservation rules of the pipe section, a digital twin system based on the pipe section's thermodynamic and fluid dynamic models can be constructed. This system receives the filtered temperature, pressure, and flow data and uses Newton's iteration method or gradient descent method, combined with the pipe section's energy and mass conservation equations, to optimize preset target parameters such as pipe section inner wall roughness, heat transfer coefficient, or local resistance coefficient online. For example, if there is a significant deviation between the model-predicted outlet temperature and the actual measured value, the system will adjust the heat transfer coefficient until the model output converges with the actual observation value, thus completing parameter verification. When calculating the physical properties of coolant in a single pipe section based on the verified target parameters, the verified pipe section parameters can be used in conjunction with a coolant property database. Through interpolation or table lookup, the average temperature, average pressure, average density, and average flow velocity of the coolant within that pipe section can be accurately calculated. For example, after obtaining the verified heat transfer coefficient and pipe section geometric parameters, the coolant temperature at the pipe section outlet can be calculated using the energy balance equation based on the coolant state at the pipe section inlet and the pipe wall temperature, thereby deducing the average physical properties of that pipe section.

[0077] Through the above technical solution, this embodiment can solve the problem of insufficient accuracy in calculating the physical properties of coolant in a single pipe section due to noise or errors in the actual collected physical property data and the uncertainty of pipe section parameters. By filtering the physical property data, noise and interference can be removed, improving data reliability. By verifying the target parameters based on the filtered physical property data and the energy and mass conservation rules of the pipe section, the uncertainty of fluid path parameters or heat exchange characteristic parameters can be corrected, thus improving the accuracy of the parameters used for calculation. Finally, by calculating the physical properties of coolant in a single pipe section based on the verified target parameters, the accuracy of the calculation results can be improved, thereby providing more reliable physical property data for subsequent leak detection and significantly improving the accuracy and precision of leak detection.

[0078] In some embodiments, in step S403, calculating the physical properties of the coolant in a single pipe section based on the verified target parameters may include, but is not limited to, the following steps: Obtain the thermodynamic property parameters of the coolant and the heat transfer coefficient of the pipe section; Calculate the physical properties of the coolant in a single pipe section based on the thermodynamic property parameter table of the coolant, the heat transfer coefficient of the pipe section, and the verified target parameters.

[0079] In some embodiments, considering only fluid path, heat exchange characteristics, and physical property data, the accuracy of the calculated physical properties of the coolant in a single pipe section cannot be guaranteed, as the thermodynamic properties of the coolant itself and the heat transfer coefficient of the pipe section also significantly affect the physical properties of the coolant. Therefore, a table of thermodynamic property parameters of the coolant and the heat transfer coefficient of the pipe section can be obtained first. Then, based on the table of thermodynamic property parameters of the coolant, the heat transfer coefficient of the pipe section, and the calibrated target parameters, the physical properties of the coolant in a single pipe section can be calculated. The table of thermodynamic property parameters of the coolant contains various thermodynamic parameters of the coolant at different temperatures and pressures, such as density, specific heat capacity, and thermal conductivity. These parameters reflect the changes in the physical properties of the coolant with temperature and pressure. The heat transfer coefficient of the pipe section reflects the heat transfer capacity of the pipe section, that is, the amount of heat transferred per unit area of ​​the pipe section per unit time.

[0080] Understandably, a coolant thermodynamic property parameter table refers to a data set recording the values ​​of physical quantities such as density, specific heat capacity, thermal conductivity, viscosity, and sound velocity of the coolant under different temperature, pressure, or concentration conditions. This data can be obtained using pre-measured and stored lookup tables, empirical formulas, or equations of state, and its purpose is to provide accurate physical property information of the coolant under various operating conditions. The heat transfer coefficient of a pipe section refers to a physical quantity characterizing the efficiency of heat transfer between the pipe section material and its inner and outer surfaces and the fluid. This coefficient can be obtained through experimental measurement, numerical simulation, or theoretical calculations based on material properties and geometric dimensions, and its purpose is to quantify the heat transfer capacity of the pipe section during heat exchange.

[0081] By employing the aforementioned technical solution, the calculation of the physical properties of the coolant in a single pipe section can fully consider the thermodynamic characteristics of the coolant itself and the heat transfer capacity of the pipe section, thereby significantly improving the accuracy of the calculation results. This accuracy provides more reliable basic data for subsequent calculations of the coolant sound velocity in the cooling pipes, thus enhancing the precision of leak point location in the overall liquid-cooled radiator coolant leak detection method. It effectively avoids misjudgments caused by deviations in physical property calculations, ensuring a rapid and accurate system response to even minor leaks.

[0082] The beneficial effects of implementing the embodiments of the present invention include: First, the pressure reading of the pressure measuring device is obtained. Based on the pressure reading, the pressure change rate is calculated using differential operation. Then, based on the pressure change rate, the start time point of the leakage event is identified. Next, the inherent response delay of the pressure measuring device is calculated. Based on the inherent response delay, the occurrence sequence of the leakage event is compensated to obtain the compensated sequence. Finally, based on the start time point and the compensated sequence, the location of the target leakage point is determined. Thus, the coolant leakage detection can be achieved by combining the pressure change rate and the inherent response delay, thereby improving the accuracy.

[0083] like Figure 2 As shown, this embodiment of the invention also provides a liquid-cooled radiator coolant leakage detection system, including: Pressure data acquisition module 501 is used to acquire pressure readings from pressure measuring equipment; Rate calculation module 502 is used to calculate the rate of pressure change based on the pressure reading using differential calculation; The time recording module 503 is used to identify the start time of a leakage event based on the rate of pressure change; The response delay quantization module 504 is used to calculate the inherent response delay of the pressure measuring device. The timing compensation module 505 is used to compensate for the occurrence timing of leakage events based on the inherent response delay, so as to obtain the compensated timing. The leak location module 506 is used to determine the location of the target leak point based on the start time point and compensation sequence.

[0084] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0085] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

Claims

1. A method for detecting coolant leakage in a liquid-cooled radiator, characterized in that, Includes the following steps: Obtain pressure readings from pressure measuring equipment; Based on the pressure readings, the rate of pressure change is calculated using differential calculations. Based on the pressure change rate, identify the start time of the leak event; Calculate the inherent response delay of the pressure measuring device; Based on the inherent response delay, the timing of the leakage event is compensated to obtain the compensated timing. The location of the target leak point is determined based on the starting time point and the compensation sequence.

2. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 1, characterized in that, The step of calculating the rate of pressure change using differential calculation based on the pressure reading includes: Obtain historical stress data; The historical pressure data and the pressure readings are subjected to feature analysis to obtain fluctuation characteristics, which include fluctuation amplitude or frequency distribution. Based on the fluctuation characteristics, the time interval for differential operations is determined; The pressure change rate is calculated based on the historical pressure data, the pressure readings, and the differential calculation time interval.

3. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 1, characterized in that, The step of identifying the start time of the leakage event based on the pressure change rate includes: If the rate of pressure change is greater than a preset threshold, a leakage signal is obtained; Local waveform analysis was performed on the leakage signal to obtain the local waveform of the signal; Within a preset time window, target points in the local waveform of the signal are identified, including waveform inflection points or local extreme points. The timestamp corresponding to the target point is used as the starting time point.

4. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 1, characterized in that, The calculation of the inherent response delay of the pressure measuring device includes: Acquire reference pressure waveform data segments and pressure waveform data streams to be matched; Perform cross-correlation calculation on the reference pressure waveform data segment and the pressure waveform data stream to be matched to obtain the cross-correlation function; The actual propagation time difference is determined based on the time offset corresponding to the peak point with the largest amplitude in the cross-correlation function. Calculate the theoretical propagation time difference between adjacent pressure measuring devices based on the physical distance between them and the sound velocity of the coolant. The inherent response delay is calculated based on the actual propagation time difference and the theoretical propagation time difference.

5. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 1, characterized in that, Determining the location of the target leak point based on the start time point and the compensation timing includes: Obtain the physical properties of the coolant in the cooling pipes; Calculate the speed of sound of the coolant in the cooling pipes based on the physical properties of the coolant. Set multiple candidate leak point locations; Based on the sound velocity of the coolant, calculate the theoretical propagation time of the pressure wave from the candidate leak point to the pressure measuring device; Based on the starting time point and the pressure change rate, the location of the first leak point is identified; Based on the theoretical propagation time and the compensation timing, identify the location of the second leakage point; The location of the target leak point is determined based on the location of the first leak point and the location of the second leak point.

6. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 5, characterized in that, The step of identifying the location of the first leak point based on the starting time point and the pressure change rate includes: The absolute value of the rate of pressure change is taken as the event intensity; Spatiotemporal correlation analysis was performed on multiple starting time points to obtain multiple pressure wave propagation paths; Identify the trend of event intensity variation in each of the pressure wave propagation paths; If the trend of change is downward, then the location of the first leak point is identified based on the starting time point and the laws of fluid physics propagation.

7. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 5, characterized in that, The acquisition of the physical properties of the coolant in the cooling pipes includes: The cooling pipes are divided into multiple pipe sections; Collect physical property data for each pipe segment; Calculate the physical properties of the coolant in a single pipe section based on the fluid path, heat exchange characteristics, and physical property data of the pipe section. The physical properties of the coolant in the cooling pipeline are obtained by performing time-smoothing processing on the physical properties of the coolant in multiple single-pipe sections.

8. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 7, characterized in that, The calculation of the physical properties of the coolant in a single pipe section based on the fluid path, heat exchange characteristics, and physical property data of the pipe section includes: The physical property data are then filtered. Based on the filtered physical property data and the energy and mass conservation rules of the pipe section, the target parameters are verified. The target parameters include the parameters of the fluid path or the parameters of the heat exchange characteristics. Based on the verified target parameters, calculate the physical properties of the coolant in the single pipe section.

9. The method for detecting coolant leakage in a liquid-cooled radiator according to claim 8, characterized in that, The step of calculating the physical properties of the coolant in the single pipe section based on the verified target parameters includes: Obtain the thermodynamic property parameters of the coolant and the heat transfer coefficient of the pipe section; The physical properties of the coolant in a single pipe section are calculated based on the thermodynamic property parameter table of the coolant, the heat transfer coefficient of the pipe section, and the verified target parameters.

10. A liquid-cooled radiator coolant leakage detection system, characterized in that, include: The pressure data acquisition module is used to acquire pressure readings from pressure measuring equipment. The rate calculation module is used to calculate the rate of pressure change based on the pressure reading using differential operations; A time recording module is used to identify the start time of a leakage event based on the pressure change rate; The response delay quantization module is used to calculate the inherent response delay of the pressure measuring device; The timing compensation module is used to compensate for the occurrence timing of the leakage event based on the inherent response delay, so as to obtain the compensated timing. The leak location module is used to determine the location of the target leak point based on the start time point and the compensation timing.

Images