High-temperature and high-pressure steam safety valve leakage detection method and system

Abstract

The invention provides a high-temperature and high-pressure steam safety valve leakage detection method and system, and relates to the technical field of industrial equipment state monitoring. Suppression processing of a first processing path is carried out on original acoustic information to generate target acoustic information, and the state of a valve is reported; and performing point-by-point subtraction on the original acoustic information and the target acoustic information through the second processing path to generate residual acoustic information. The differential processing can effectively strip real and weak continuous leakage signals which may be weakened together by a conventional noise suppression strategy. And then, analyzing the residual acoustic information, and judging whether continuous acoustic information meeting a preset duration threshold and a preset energy threshold exists or not, thereby accurately identifying a real leakage signal. And if the continuous real leakage signal exists, an early warning record is triggered.

Description

Technical Field

[0001] The present application relates to the technical field of industrial equipment status monitoring, and in particular to a high-temperature and high-pressure steam safety valve leakage detection method and system. Background Art

[0002] In industrial production environments, particularly in high-temperature, high-pressure steam systems in power plants, safety valves are critical safety equipment, and their sealing integrity is crucial to production stability and personnel safety. Therefore, timely and accurate detection of steam leaks from safety valves is extremely important. Existing leak detection systems typically rely on acoustic information collection and analysis. However, in actual operation, especially under abnormal operating conditions such as unit power adjustments, complex environmental factors and human intervention often introduce interference, making it difficult for the system to accurately identify weak leak signals, and even resulting in false alarms or missed detections.

[0003] Specifically, when the high-temperature, high-pressure steam safety valve is adjusting the unit power, water droplets may condense on the valve body surface due to the drop in temperature. The pulsed sound produced by the water droplets hitting the valve body may have acoustic characteristics similar to those of a real steam leak signal, causing the system to mistakenly identify a leak. To further complicate matters, to suppress such false alarms, operators may temporarily enable the environmental noise suppression mode. While this mode filters out interfering noise, it may also accidentally weaken the real, weak, and continuous steam leak signal, causing it to fall below the alarm threshold, resulting in the system being unable to identify the real leak and causing a missed alarm. This dilemma of coexisting false alarms and missed alarms seriously affects the reliability of leak detection and the operational safety of the equipment. Summary of the Invention

[0004] The present application provides a high-temperature and high-pressure steam safety valve leakage detection method and system, which aims to solve the technical problems in the prior art that, in complex industrial environments, the high-temperature and high-pressure steam safety valve leakage detection system is easily interfered by environmental noise and human operation, resulting in difficulty in accurately identifying weak real leakage signals, and the coexistence of false alarms and missed alarms.

[0005] In one aspect, the present application provides a method for detecting leakage of a high-temperature and high-pressure steam safety valve, comprising: receiving an ambient noise suppression mode activation instruction from an operator; In response to the instruction to enable the ambient noise suppression mode, performing suppression processing of the first processing path on the raw acoustic information collected in real time to generate target acoustic information, and reporting the valve status of the high-temperature and high-pressure steam safety valve based on the target acoustic information; subtracting the original acoustic information from the target acoustic information point by point through a second processing path to generate residual acoustic information; Analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold, so as to determine whether there is a persistent real leakage signal; If it is determined that the persistent real leakage signal exists, an early warning record is triggered.

[0006] Optionally, the step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal further includes: If there is continuous acoustic information that meets a preset duration threshold and a preset energy threshold, extracting an instantaneous energy envelope curve from the continuous acoustic information; Performing spectrum analysis on the instantaneous energy envelope curve to obtain spectrum distribution information; Determine whether there are periodic peaks in the low-frequency band spectrum in the spectrum distribution information. If there are one or more periodic peaks higher than a preset threshold, and the frequency of the periodic peaks is consistent with the dripping frequency range of water droplets, it is determined that there is no persistent real leakage signal. If there are no periodic peaks, it is determined that there is a persistent real leakage signal.

[0007] Optionally, the step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal includes: If there is continuous acoustic information that meets a preset duration threshold and a preset energy threshold, accurately aligning the energy change envelope of the continuous acoustic information with the amplitude change envelope of the vibration signal collected by one or more mechanical vibration sensors preset in the steam pipe system in time; Within a preset analysis period, the correlation coefficient between the energy change envelope of the continuous acoustic information and the amplitude change envelope of the vibration signal is calculated; if the correlation coefficient is higher than the preset correlation coefficient threshold, it is determined that there is no continuous real leakage signal; if the correlation coefficient is lower than the preset correlation coefficient threshold, it is determined that there is a continuous real leakage signal.

[0008] Optionally, the step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold includes: Calculating the short-time energy of the residual acoustic information; Counting the distribution of the short-term energy within a preset duration threshold; determining a background noise level of the residual acoustic information based on a preset low-energy portion of the distribution of the short-term energy; adjusting a preset energy threshold of the continuous acoustic information according to the background noise level to obtain a target energy threshold; It is determined whether the short-term energy is within a preset duration threshold and is maintained above the target energy threshold.

[0009] Optionally, the step of determining the background noise level of the residual acoustic information based on a preset low-energy portion of the distribution of the short-time energy includes: identifying a set of low-energy points from the short-time energies; Performing stability analysis on the low-energy point set to obtain a stability analysis result; Based on the stability analysis result, a background noise level of the residual acoustic information is determined.

[0010] Optionally, the step of performing stability analysis on the low-energy point set to obtain a stability analysis result includes: Calculating the dispersion of energy distribution of the low-energy point set within a preset duration threshold; The stability of the low-energy point set is determined based on the discreteness of the energy distribution to obtain a stability analysis result.

[0011] Optionally, the step of calculating the dispersion of energy distribution of the low-energy point set within a preset duration threshold includes: Obtaining energy values ​​of the low-energy point set within the preset duration threshold; Based on the energy value, calculating the interquartile range of the energy value; or, based on the energy values, calculating the median absolute deviation of the energy values; The interquartile range or the median absolute deviation is used as the dispersion of the energy distribution.

[0012] Optionally, the step of judging the stability of the low-energy point set according to the discreteness of the energy distribution includes: comparing the dispersion of the energy distribution with a preset stability threshold; If the dispersion is lower than the preset stability threshold, the low energy point set is determined to be stable.

[0013] Optionally, the step of determining whether there are periodic peaks in the low-frequency spectrum in the spectrum distribution information, and determining that there is no persistent real leakage signal if there are one or more periodic peaks higher than a preset threshold and the frequencies of the periodic peaks are consistent with the dripping frequency range of water droplets, and if there are no periodic peaks, determining that there is a persistent real leakage signal includes: Determining a preset frequency interval corresponding to a frequency range of water droplets from a low-frequency frequency band in the frequency spectrum distribution information; Calculating the total spectrum energy within the preset frequency interval; The sum of the spectrum energy is compared with a preset threshold. If the sum of the spectrum energy exceeds the preset threshold, it is determined that there is no persistent real leakage signal. If the sum of the spectrum energy does not exceed the preset threshold, it is determined that there is a persistent real leakage signal.

[0014] On the other hand, the present application provides a high-temperature and high-pressure steam safety valve leakage detection system, comprising: A receiving module, configured to receive an instruction from an operator to enable an ambient noise suppression mode; a first path processing module, configured to, in response to the ambient noise suppression mode activation instruction, perform a first processing path suppression process on the raw acoustic information collected in real time to generate target acoustic information, and report the valve status of the high-temperature and high-pressure steam safety valve based on the target acoustic information; a second path processing module, configured to subtract the original acoustic information from the target acoustic information point by point through a second processing path to generate residual acoustic information; An independent analysis module is used to analyze whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold, so as to determine whether there is a persistent real leakage signal; The early warning recording module is used to trigger the early warning recording if it is determined that the persistent real leakage signal exists.

[0015] The present application provides a method and system for detecting leakage of a high-temperature and high-pressure steam safety valve. Through a dual-path processing mechanism, the accuracy and reliability of leakage detection of a high-temperature and high-pressure steam safety valve are significantly improved. The first path suppresses environmental noise and reports the valve status. The second path removes the weak continuous leakage signal that is conventionally suppressed through residual analysis, effectively solving the problem that the real leakage signal is easily mistakenly filtered out in a complex industrial environment, and the missed alarm rate is greatly reduced. At the same time, based on the preset duration and energy threshold, the residual signal is analyzed to accurately distinguish between instantaneous interference and real leakage, and the false alarm rate is significantly reduced. The dual paths work together to achieve high-sensitivity detection of weak leaks while ensuring conventional monitoring functions, adapting to strong noise conditions such as high temperature and high pressure, improving system robustness, ensuring that the early warning is true and reliable, and the overall technical effect is outstanding. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0017] Figure 1Schematic diagram of a flow chart of a high-temperature and high-pressure steam safety valve leakage detection method in an embodiment is shown in FIG. Figure 2 Schematic diagram of a module configuration of a high-temperature and high-pressure steam safety valve leakage detection system in an embodiment is shown in FIG.

[0018] Figure numerals: 100, high-temperature and high-pressure steam safety valve leakage detection system; 10, receiving module; 20, first path processing module; 30, second path processing module; 40, independent analysis module; 50, early warning recording module. DETAILED DESCRIPTION

[0019] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0020] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0021] In industrial production environments, particularly in high-temperature, high-pressure steam systems in power plants, safety valves are critical safety devices whose sealing integrity is crucial to production stability and personnel safety. Therefore, timely and accurate detection of steam leaks from safety valves is extremely important. Existing leak detection systems typically rely on acoustic information collection and analysis. However, in actual operation, especially under abnormal operating conditions such as unit power adjustments, complex environmental factors and human intervention often introduce interference, making it difficult for the system to accurately identify actual weak leak signals, and even resulting in false alarms or missed alarms. For example, when water droplets condense and drip from the valve body, the pulsed sound they produce can be similar to a true steam leak signal, leading to system misjudgment. Furthermore, the ambient noise suppression mode activated to suppress such false alarms may, while filtering out interference, inadvertently attenuate the actual weak, continuous steam leak signal, dropping it below the alarm threshold, resulting in missed alarms. This dilemma of both false alarms and missed alarms seriously impacts the reliability of leak detection and the operational safety of the equipment.

[0022] like Figure 1FIG2 is a schematic diagram showing a method for detecting leakage of a high-temperature and high-pressure steam safety valve. The present application proposes a method for detecting leakage of a high-temperature and high-pressure steam safety valve, comprising: S10, receiving an instruction from an operator to enable an ambient noise suppression mode.

[0023] The ambient noise suppression mode activation command refers to a command issued by an operator or an automated system to activate a specific noise suppression function in the acoustic information processing system to cope with a high noise environment.

[0024] S20, in response to the environmental noise suppression mode activation instruction, performing suppression processing of the first processing path on the original acoustic information collected in real time to generate target acoustic information, and reporting the valve status of the high-temperature and high-pressure steam safety valve based on the target acoustic information.

[0025] Raw acoustic information refers to the unprocessed raw sound data stream collected in real time by acoustic sensors (e.g., microphones, ultrasonic sensors, etc.). It includes all acoustic components such as ambient noise, equipment operation noise, and possible leakage signals.

[0026] The first processing path is a conventional path that performs noise suppression on the raw acoustic information, generating clear acoustic information for routine condition monitoring. This path typically employs various filtering and noise reduction algorithms, such as adaptive noise cancellation, spectral subtraction, and wavelet denoising, to effectively remove ambient background noise.

[0027] The target acoustic information is the acoustic data obtained after the suppression processing of the first processing path, which is mainly used for conventional valve status monitoring and reporting, such as determining whether the valve is in an open, closed or abnormal state.

[0028] S30 , subtracting the original acoustic information from the target acoustic information point by point through a second processing path to generate residual acoustic information.

[0029] The second processing path is another processing path that is parallel or in series with the first processing path. Its core is to extract acoustic components with potential leakage characteristics that may be suppressed or weakened in the first processing path by subtracting the original acoustic information from the target acoustic information point by point.

[0030] The residual acoustic information is the difference between the original and target acoustic information, obtained by point-by-point subtraction. In theory, if the first processing path perfectly suppresses all non-leakage noise, the residual acoustic information will primarily contain suppressed leakage signals or incompletely suppressed noise.

[0031] S40: Analyze whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold, so as to determine whether there is a persistent real leakage signal.

[0032] The preset duration threshold and energy threshold are key parameters for determining whether acoustic information represents a persistent leakage signal. The duration threshold distinguishes transient noise from persistent signals, while the energy threshold distinguishes weak signals from significant ones. These thresholds can be preset and adjusted based on actual operating conditions, equipment type, and experience.

[0033] A true leakage signal refers to an acoustic signal generated by the leakage of high-temperature and high-pressure steam with a certain duration and energy characteristics. It is different from instantaneous interference in the environment or non-leakage noise such as dripping water.

[0034] S50: If it is determined that the persistent real leakage signal exists, triggering an early warning record.

[0035] An early warning record is an alarm, log record or operator notification triggered when the system determines that there is a persistent real leakage signal, prompting further inspection and processing.

[0036] This application introduces a dual-path processing mechanism. Specifically, in ambient noise suppression mode, a first processing path generates target acoustic information for valve status reporting. A second processing path then subtracts the original acoustic information from the target acoustic information point by point to generate residual acoustic information. This allows for effective identification and determination of persistent true leak signals from the residual acoustic information, avoiding the false positives and false negatives inherent in traditional methods and significantly improving the accuracy and reliability of leak detection.

[0037] The high-temperature, high-pressure steam safety valve leak detection method disclosed in this application aims to address the challenges of safety valve leak detection in complex industrial environments, particularly in high-temperature, high-pressure steam systems. This method is primarily applicable to industrial scenarios with high-temperature, high-pressure steam piping systems, such as power plants and chemical plants.

[0038] The high-temperature and high-pressure steam safety valve leakage detection method disclosed in this application may include the following steps: First, the system needs to receive an operator-initiated command to activate ambient noise suppression mode. This command can be issued in a variety of ways. For example, the operator can manually trigger the command using a physical button on the human-machine interface or a virtual button on the touchscreen. Alternatively, the command can be sent remotely from a control system, such as a central control room connected via a network. Furthermore, the system can be triggered by pre-set automated rules, such as automatically generating and issuing the command when the system detects that the ambient noise level exceeds a certain threshold.

[0039] Next, in response to the above-mentioned instruction to activate the ambient noise suppression mode, the real-time raw acoustic information is subjected to suppression processing using the first processing path to generate target acoustic information. Based on this target acoustic information, the valve status of the high-temperature, high-pressure steam safety valve is reported. Specifically, the raw acoustic information can be collected in real time by one or more acoustic sensors positioned near the safety valve. The suppression process in the first processing path can employ a variety of noise suppression algorithms. For example, a spectral subtraction-based approach can be employed to estimate the spectrum of background noise and subtract it from the spectrum of the raw acoustic information to suppress ambient noise. As another embodiment, adaptive filtering technology can be employed to estimate and eliminate noise components using an adaptive filter. Alternatively, a wavelet transform denoising method can be employed to perform wavelet decomposition on the raw acoustic information, followed by thresholding the wavelet coefficients and then performing wavelet reconstruction to achieve denoising. The generated target acoustic information is then used for regular valve status reporting. For example, by analyzing the energy and frequency characteristics of the target acoustic information, it can be determined whether the valve is fully closed, slightly open, or fully open. This status information is then reported to a monitoring center via a data bus or wireless communication.

[0040] Furthermore, the original acoustic information is subtracted from the target acoustic information point by point through the second processing path to generate residual acoustic information. This step is one of the key aspects of this application. For example, in the first processing path, certain weak, persistent leakage signals may be mistaken for noise and suppressed. By performing point-by-point time or frequency domain subtraction between the original acoustic information (including all information) and the target acoustic information (information that has undergone suppression processing), the signal components that were suppressed or weakened in the first processing path can be effectively extracted. Specifically, the original acoustic information and the target acoustic information can be precisely aligned on the time axis, and then a numerical subtraction operation is performed on each sampling point. For example, if the sampling value of the original acoustic information at a certain moment is A, and the sampling value of the target acoustic information at the same moment is B, then the sampling value of the residual acoustic information at that moment is AB. This point-by-point subtraction operation allows the signals that were "filtered out" or "weakened" in the first processing path to be revealed, thereby forming residual acoustic information.

[0041] Furthermore, it is necessary to analyze whether there is continuous acoustic information in the above-mentioned residual acoustic information that meets the preset duration threshold and the preset energy threshold to determine whether there is a continuous real leakage signal. The analysis process may include multiple steps. For example, the short-term energy of the residual acoustic information can be first calculated to obtain the change of its energy over time. Then, through a sliding window or other methods, it is statistically analyzed whether the energy of the residual acoustic information continues to remain above the preset energy threshold within the preset duration threshold. For example, if the energy of the residual acoustic information is higher than a certain decibel value (preset energy threshold) for 5 consecutive seconds (preset duration threshold), it is preliminarily determined that there is continuous acoustic information. This step is intended to distinguish between instantaneous noise (such as water dripping sound) and continuous leakage signals.

[0042] Finally, if a persistent, real leak signal is detected, an early warning is triggered. If the above analysis confirms a persistent, real leak, the system initiates an early warning. For example, an audible and visual alarm may be issued to the operator, and an alert message may be displayed on the monitoring interface. The system also records relevant data, such as the leak event time, duration, and energy level, in a database for subsequent troubleshooting and maintenance. Furthermore, the early warning information can be sent to the relevant personnel via text message, email, or other means to ensure a timely response.

[0043] The high-temperature and high-pressure steam safety valve leakage detection method disclosed in this application has an overall working principle of effectively solving the challenges faced by traditional methods in complex industrial environments, especially in safety valve leakage detection in high-temperature and high-pressure steam systems, through an innovative dual-path processing mechanism.

[0044] Specifically, when the operator issues the command to enable ambient noise suppression mode, the system simultaneously initiates two parallel acoustic information processing paths. The first processing path performs conventional noise suppression on the raw acoustic information collected in real time, generating the target acoustic information used for valve status reporting. This path ensures that the basic operating status of the valve, such as the valve's open or closed state, can be accurately monitored and reported even in noisy environments. However, traditional methods that rely solely on this path can result in weak true leakage signals being filtered out or weakened during the noise suppression process, resulting in missed detections.

[0045] To address this issue, the present application introduces a second processing path. This path performs a point-by-point subtraction between the unprocessed original acoustic information and the target acoustic information processed by the first processing path. This ingenious design allows the signal components that are suppressed or weakened in the first processing path but have potential leakage characteristics to appear, thereby generating residual acoustic information. The residual acoustic information essentially represents the part of the original signal that is "removed" by the first processing path. If the first processing path mainly removes the ambient noise, then the residual acoustic information may contain real leakage signals that are mistakenly judged as noise.

[0046] The system then conducts an in-depth analysis of the residual acoustic information to determine whether it contains persistent acoustic signals that meet preset duration and energy thresholds. This analysis step is crucial for distinguishing true leak signals from transient interference, such as the sound of dripping water. Water dripping sounds are typically pulsed and short-lived, while steam leak signals are persistent. By setting appropriate duration and energy thresholds, the system can effectively filter out transient, low-energy interference signals, thereby focusing on the persistent, high-energy signals that may represent true leaks.

[0047] Therefore, when the above analysis confirms the existence of continuous acoustic information that meets the conditions, the system determines that there is a continuous real leakage signal and immediately triggers the early warning record. This mechanism ensures that even in the environmental noise suppression mode, weak real leakage signals will not be missed, and can effectively avoid false alarms caused by non-leakage noise such as water dripping. Through the close coordination of the above steps, the method of this application can significantly improve the accuracy and reliability of high-temperature and high-pressure steam safety valve leakage detection, thereby ensuring the safe and stable operation of industrial production.

[0048] The core innovation of the high-temperature and high-pressure steam safety valve leakage detection method disclosed in this application lies in the introduction of a dual-path acoustic information processing mechanism, combined with the analysis of residual acoustic information, to solve the dilemma of false alarms and missed alarms coexisting in traditional methods in complex industrial environments, especially in the presence of water droplet sound interference and environmental noise suppression mode.

[0049] Compared to existing leak detection methods that rely solely on a single noise suppression path, this application offers the advantage that, when using ambient noise suppression mode, traditional methods often indiscriminately filter out all signals identified as noise. This can include weak but real, persistent steam leak signals, leading to missed alarms. For example, when an operator activates noise suppression mode to suppress transient interference like dripping water, the actual steam leak signal may be weakened due to its lower energy or similar frequency characteristics to the noise, ultimately failing to trigger an alarm.

[0050] The present application introduces a second processing path to perform point-by-point subtraction between the original acoustic information and the target acoustic information after the suppression processing by the first processing path, thereby generating residual acoustic information. This innovative step allows potential leakage signals that are suppressed or weakened in the first processing path to be "restored" or "highlighted." By judging the duration threshold and energy threshold of the above-mentioned residual acoustic information, the present application can effectively identify real leakage signals with continuous characteristics and distinguish them from transient interference (such as the sound of water drops). The sound of water drops usually appears as a short-term pulse, while steam leaks are continuous. Therefore, even in the ambient noise suppression mode, the present application can accurately capture weak continuous leakage signals and avoid missed reports. At the same time, by analyzing the characteristics of the residual signal, false alarms caused by non-leakage noises such as water drops are also effectively avoided.

[0051] In summary, the method of the present application achieves accurate and reliable detection of high-temperature and high-pressure steam safety valve leakage in a complex noise environment through dual-path processing and residual acoustic information analysis, significantly improving the anti-interference ability and recognition accuracy of the detection system, thereby effectively ensuring the safety of industrial production and the operational stability of equipment, and reflecting significant progress.

[0052] In some embodiments, the step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal further includes: If there is continuous acoustic information that meets a preset duration threshold and a preset energy threshold, extracting an instantaneous energy envelope curve from the continuous acoustic information; Performing spectrum analysis on the instantaneous energy envelope curve to obtain spectrum distribution information; Determine whether there are periodic peaks in the low-frequency band spectrum in the spectrum distribution information. If there are one or more periodic peaks higher than a preset threshold, and the frequency of the periodic peaks is consistent with the dripping frequency range of water droplets, it is determined that there is no persistent real leakage signal. If there are no periodic peaks, it is determined that there is a persistent real leakage signal.

[0053] Specifically, when the system detects continuous acoustic information that meets a preset duration threshold and a preset energy threshold, in order to further accurately determine whether it is a real leakage signal, it is necessary to perform a deeper analysis of the continuous acoustic information.

[0054] The energy intensity profile of the acoustic signal that changes over time is obtained through signal processing techniques, such as the Hilbert transform or the sliding average method, so that an instantaneous energy envelope curve can be extracted from the persistent acoustic information. This instantaneous energy envelope curve can reflect the instantaneous amplitude or energy change trend of the acoustic signal.

[0055] The extracted instantaneous energy envelope curve is subjected to spectral analysis, such as Fourier transform, to reveal its energy distribution at different frequencies. This step aims to identify the presence of specific frequency components, particularly periodic components, within the energy envelope. The spectral distribution information is then used to determine whether the low-frequency spectrum in the spectrum contains periodic peaks. If one or more periodic peaks are present that exceed a preset threshold, and the frequency of these periodic peaks matches the droplet frequency range, then a persistent true leak signal is determined to be absent. If no periodic peaks are present, then a persistent true leak signal is determined to be present.

[0056] The low-frequency spectrum typically ranges from 0 Hz to several hundred Hz. The frequency range of falling water droplets typically exhibits specific low-frequency periodic characteristics, such as between a few Hz and tens of Hz. By detecting these specific periodic peaks, the acoustic signal caused by falling water droplets can be effectively identified, distinguishing it from a true steam leak signal.

[0057] The technical solution of this application solves the problem of traditional methods misidentifying dripping water as steam leaks by performing spectral analysis on the instantaneous energy envelope of continuous acoustic information, paying special attention to whether there are periodic peaks in the low-frequency spectrum that match the frequency range of dripping water. Steam leaks typically produce broadband continuous noise, whose energy envelope generally lacks obvious low-frequency periodicity. However, dripping water produces a series of discrete impact sounds, the repetition of which causes its instantaneous energy envelope to exhibit obvious periodicity in the low-frequency band. Therefore, by identifying this unique periodic feature, interference from non-leakage sources can be effectively eliminated, improving the accuracy of leak detection.

[0058] Through the above technical solution, this application can significantly improve the accuracy of leak detection for high-temperature, high-pressure steam safety valves, effectively avoiding false alarms caused by non-leakage factors such as dripping water. This not only reduces unnecessary maintenance inspections and wastes resources, but also ensures that early warnings are triggered in the event of a leak, thereby improving the safety and reliability of equipment operation.

[0059] For example, consider a high-temperature, high-pressure steam pipeline system. As a result of condensation, water droplets occasionally drip onto valves or pipe walls, producing a continuous "tick-tick" sound. This sound may meet preset duration and energy thresholds, leading to it being mistakenly identified as a steam leak in systems that rely solely on these thresholds. However, according to the technical solution of this application, when the system detects this persistent acoustic signal, it further extracts its instantaneous energy envelope. This energy envelope is then subjected to spectral analysis. If the analysis results reveal one or more distinct periodic peaks in the low-frequency range (e.g., 5Hz to 20Hz), and the frequencies of these peaks match the known droplet frequency range, the system will determine that there is no persistent, true leak signal. Conversely, if the spectral analysis results do not detect such periodic peaks in the low-frequency range, the system will determine that there is a persistent, true leak signal. For example, when a steam leak actually occurs, the energy envelope of its acoustic information will exhibit relatively stable or random characteristics, without distinct periodic peaks in the low-frequency range, thus being correctly identified as a leak. In this way, the technical solution of the present application can effectively distinguish between real steam leaks and non-leakage phenomena such as water dripping, greatly reducing the false alarm rate and improving the reliability of the leak detection system.

[0060] In some embodiments, the step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal includes: If there is continuous acoustic information that meets a preset duration threshold and a preset energy threshold, accurately aligning the energy change envelope of the continuous acoustic information with the amplitude change envelope of the vibration signal collected by one or more mechanical vibration sensors preset in the steam pipe system in time; Within a preset analysis period, the correlation coefficient between the energy change envelope of the continuous acoustic information and the amplitude change envelope of the vibration signal is calculated; if the correlation coefficient is higher than the preset correlation coefficient threshold, it is determined that there is no continuous real leakage signal; if the correlation coefficient is lower than the preset correlation coefficient threshold, it is determined that there is a continuous real leakage signal.

[0061] Specifically, when the system identifies continuous acoustic information that meets a preset duration threshold and a preset energy threshold, additional judgment basis needs to be introduced to further verify whether it is a real leakage signal. Among them, the energy intensity change trend of the acoustic signal in the time dimension can be obtained by performing short-time energy calculation or envelope extraction algorithm on the acoustic signal. The mechanical vibration sensor can be arranged near the safety valve, on the pipe wall or other equipment that may generate vibration, and is used to collect the mechanical vibration signal of the system in real time. Accurately aligning the energy change envelope of the acoustic information with the amplitude change envelope of the vibration signal in time means that it is necessary to ensure that the two are synchronized on the time axis for effective comparative analysis, which can be achieved through timestamp synchronization, signal resampling or cross-correlation techniques.

[0062] Furthermore, within a preset analysis period, a correlation coefficient is calculated between the energy variation envelope of the persistent acoustic information and the amplitude variation envelope of the vibration signal. The preset analysis period can be set based on the actual application scenario and signal characteristics, for example, from a few seconds to tens of seconds. The correlation coefficient quantifies the strength and direction of the linear relationship between two signals and can be, for example, a Pearson correlation coefficient or a cross-correlation function. If the calculated correlation coefficient is higher than a preset correlation coefficient threshold, it indicates a high degree of synchronization or a causal relationship between the acoustic signal and the mechanical vibration signal. In this case, it can be determined that the persistent acoustic information is not caused by a true steam leak, but rather by mechanical vibration, and therefore, a persistent true leak signal is not present. Conversely, if the correlation coefficient is lower than the preset correlation coefficient threshold, it indicates a weak correlation between the acoustic signal and the mechanical vibration signal. In this case, it can be determined that the persistent acoustic information is more likely caused by a true steam leak, and therefore, a persistent true leak signal is present. The preset correlation coefficient threshold can be set based on historical data, experimental results, or expert experience to balance the false positive rate and false negative rate.

[0063] The technical solution of this application effectively addresses the problem of misjudgment that can result from relying solely on acoustic information by introducing mechanical vibration signals as an auxiliary basis for judgment. When mechanical vibration occurs in a steam pipe system, it typically generates both a vibration signal that can be captured by a mechanical vibration sensor and an acoustic signal that can be captured by an acoustic sensor. These acoustic and vibration signals caused by mechanical vibration often have a high temporal correlation, meaning their energy or amplitude variation trends are synchronized. By precisely aligning the energy variation envelope of the persistent acoustic information with the amplitude variation envelope of the mechanical vibration signal and calculating their correlation coefficient, acoustic signals that are highly correlated with mechanical vibration can be identified. A high correlation coefficient indicates that the acoustic signal is likely due to mechanical vibration rather than a steam leak, thus eliminating it. Actual steam leak signals, on the other hand, are typically generated by fluid passing through small holes or cracks. Their acoustic characteristics have a low correlation with mechanical vibration signals, resulting in a low correlation coefficient between their energy variation envelope and the amplitude variation envelope of the mechanical vibration signal. Therefore, this multimodal data fusion analysis can more accurately distinguish true leak signals from interference signals caused by background noise or mechanical vibration.

[0064] Through the above technical solution, the present application can significantly improve the accuracy and reliability of leakage detection of high-temperature and high-pressure steam safety valves. By introducing mechanical vibration signals as a means of cross-validation, it effectively avoids the situation where the acoustic signals caused by mechanical vibrations inside the system are misjudged as leakage signals, thereby significantly reducing the false alarm rate. This not only reduces unnecessary on-site inspections and maintenance costs, but also allows maintenance personnel to focus more on dealing with real leakage problems, improving the safety and efficiency of equipment operation. In addition, this technical solution provides a more robust leakage judgment mechanism, which is particularly suitable for scenarios with diverse noise sources and severe interference in complex industrial environments.

[0065] For example, assume that an acoustic sensor and a mechanical vibration sensor are installed near a safety valve in a high-temperature, high-pressure steam pipeline system. When a pump in the system starts or stops, it may generate continuous mechanical vibrations and a certain amount of acoustic noise. In this case, after the raw acoustic information collected by the acoustic sensor passes through the first and second processing paths, continuous acoustic information that meets preset duration and energy thresholds may be detected in the residual acoustic information.

[0066] To determine whether this is a real leak, the system further extracts the energy variation envelope of the persistent acoustic information. Simultaneously, the mechanical vibration sensor also captures the vibration signal synchronized with the pump start / stop and extracts its amplitude variation envelope. The system precisely aligns these two envelopes in time and calculates the correlation coefficient between them within a preset analysis period (e.g., 5 seconds). If the calculated correlation coefficient (e.g., 0.9) is higher than the preset correlation coefficient threshold (e.g., 0.7), the system determines that the persistent acoustic information is caused by the pump's mechanical vibration, not a real steam leak, and therefore does not trigger an alert.

[0067] On the contrary, if a small leak occurs in the safety valve, the generated steam leakage acoustic signal is detected in the residual acoustic information as continuous acoustic information that meets the duration threshold and energy threshold. At this time, since the leakage acoustic signal has no direct causal relationship with the mechanical vibration in the system, the correlation coefficient between its energy change envelope and the amplitude change envelope of the vibration signal collected by the mechanical vibration sensor will be very low (for example, 0.2). When this correlation coefficient is lower than the preset correlation coefficient threshold, the system will accurately determine the presence of a continuous real leakage signal and trigger an early warning record, thereby notifying maintenance personnel in time for processing. In this way, the technical solution of the present application can effectively distinguish between mechanical vibration noise and real leakage signals, thereby improving the accuracy of detection.

[0068] In some embodiments, the step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold includes: Calculating the short-time energy of the residual acoustic information; Counting the distribution of the short-term energy within a preset duration threshold; determining a background noise level of the residual acoustic information based on a preset low-energy portion of the distribution of the short-term energy; adjusting a preset energy threshold of the continuous acoustic information according to the background noise level to obtain a target energy threshold; It is determined whether the short-term energy is within a preset duration threshold and is maintained above the target energy threshold.

[0069] Specifically, the energy of the residual acoustic signal is calculated within a short time window to obtain the short-term energy of the residual acoustic information, which reflects the signal strength within the instantaneous time period. For example, this can be done by framing the signal and calculating the sum of squared energies or root mean square (RMS) of each frame. This is used to obtain the energy characteristics of the residual acoustic information over time, providing basic data for subsequent noise level assessment and threshold adjustment.

[0070] Among them, statistically analyzing the distribution of the short-time energy within a preset duration threshold can be understood as performing statistical analysis on the calculated short-time energy values ​​within a preset time length, such as generating an energy histogram or cumulative distribution function, which is used to understand the energy fluctuation range and main concentration area of ​​the residual acoustic information within a specific time period, which helps to identify the typical energy level of background noise.

[0071] In practical applications, the background noise level of the residual acoustic information is determined based on a preset low-energy portion of the distribution of the short-term energy. Specifically, this involves identifying an energy interval representing background noise from the statistical distribution of the short-term energy. For example, the lowest few percent (e.g., 5% or 10%) of energy values ​​in the energy distribution can be selected as representative of background noise, or low-energy clusters can be identified through methods such as cluster analysis to accurately estimate the non-leakage acoustic background in the current environment and provide a benchmark for dynamically adjusting the detection threshold.

[0072] Furthermore, the preset energy threshold for the persistent acoustic information is adjusted based on the background noise level to obtain a target energy threshold. This means that when the background noise level is high, the detection threshold is increased accordingly to avoid misidentifying noise as a leak; when the background noise level is low, the detection threshold is decreased accordingly to increase sensitivity to weak leakage signals. This ensures that the detection threshold is always adapted to the actual current environmental noise level.

[0073] Finally, a determination is made as to whether the short-term energy is within a preset duration threshold and remains above the target energy threshold. This step is based on the dynamically adjusted target energy threshold for the final leakage signal determination. A persistent true leakage signal is only determined if the short-term energy of the residual acoustic information remains above the dynamically adjusted target energy threshold for a sufficiently long period of time (meeting the preset duration threshold).

[0074] The technical solution of the present application effectively solves the problem of false alarms or missed alarms that may occur in complex and changeable noise environments with traditional fixed threshold methods by introducing dynamic evaluation of the background noise level of residual acoustic information and adaptive adjustment of the energy threshold. Specifically, first, by calculating the short-time energy and statistically analyzing its distribution, the energy characteristics of the residual acoustic information in the time dimension can be captured. Secondly, by analyzing the low-energy part in the short-time energy distribution, the current background noise level can be accurately identified and quantified. It is precisely because the background noise level can be obtained in real time that the preset energy threshold can be dynamically adjusted according to the actual environment to generate a target energy threshold that matches the current noise environment. As a result, when making the final leakage judgment, the real leakage signal and environmental noise can be distinguished more accurately, thereby significantly improving the accuracy and robustness of the detection.

[0075] Through the above-mentioned technical solution, this application can overcome the challenges faced by traditional fixed-threshold detection methods when background noise fluctuates significantly. By evaluating the background noise level of residual acoustic information in real time and dynamically adjusting the energy threshold, the system can adaptively adapt to different environmental noise conditions, thereby significantly reducing the false alarm rate and missed alarm rate. This makes leak detection of high-temperature and high-pressure steam safety valves more accurate and reliable, especially in complex and changing industrial site noise environments. It can effectively improve the performance and practicality of leak detection and ensure the safety of equipment operation.

[0076] For example, assume that during the operation of a high-temperature, high-pressure steam safety valve, the background noise level changes periodically due to the start and stop of surrounding equipment or the operation of an ambient fan. Without adopting the technical solution of the present application, if a fixed energy threshold is used, when the background noise suddenly increases, the noise component in the residual acoustic information may exceed the fixed threshold and be misjudged as a leak, resulting in unnecessary early warning. Conversely, when the background noise suddenly decreases, if the fixed threshold is set too high, it may cause the weak real leakage signal to be submerged below the threshold, resulting in failure to detect the leak in time.

[0077] However, using the technical solution of the present application, the system continuously calculates the short-term energy of the residual acoustic information and calculates its distribution within a preset duration threshold. For example, by analyzing the low-energy portion of the short-term energy distribution, the system can estimate the current background noise level in real time. When the background noise increases, the system correspondingly raises the target energy threshold, ensuring that only persistent signals with an intensity significantly higher than the current noise level are identified as leaks, effectively avoiding false alarms caused by noise. When the background noise decreases, the system correspondingly lowers the target energy threshold, increasing sensitivity to weak leakage signals and ensuring that even minor leaks are captured promptly. For example, if the system detects an average value of X for the low-energy portion of the short-term energy over a certain period of time, the target energy threshold is set to X + Y (Y is a preset safety margin). When the background noise increases and the average value of the low-energy portion reaches X' (X'>X), the target energy threshold is adjusted to X' + Y, thus achieving adaptive threshold adjustment. As a result, regardless of changes in background noise, the technical solution of the present application can provide a more accurate and reliable leak detection, significantly improving the robustness of the detection system.

[0078] In some embodiments, the step of determining the background noise level of the residual acoustic information based on the preset low-energy portion of the distribution of the short-time energy comprises: identifying a set of low-energy points from the short-time energies; Performing stability analysis on the low-energy point set to obtain a stability analysis result; Based on the stability analysis result, a background noise level of the residual acoustic information is determined.

[0079] Specifically, after performing short-time energy calculation on the residual acoustic information and counting its distribution, it is necessary to identify the low-energy part representing the background noise from the distribution of the short-time energy. This is usually done by setting an energy threshold or using a statistical method (e.g., based on the lower quartile or a certain percentile of the energy distribution) to identify those time points with lower energy values, and then collecting these points to form the set of low-energy points. The set of low-energy points is considered as potential background noise samples.

[0080] Further, in order to ensure that the identified set of low-energy points indeed represents stable background noise rather than transient energy fluctuations or outliers, it is necessary to perform stability analysis on the set of low-energy points. The stability analysis aims to evaluate whether the energy distribution of these low-energy points is sufficiently concentrated and stable within a pre-set duration threshold. For example, the stability can be measured by calculating the dispersion (such as variance, standard deviation, interquartile range, or median absolute deviation) of the energy distribution of these low-energy points. If the dispersion is small, it indicates that these low-energy points are stable and can reliably represent the background noise.

[0081] Thus, after obtaining the stability analysis result of the set of low-energy points, the background noise level of the residual acoustic information can be determined based on the result. For example, if the stability analysis result indicates that the set of low-energy points is stable, the average energy value, median energy value, or some statistical quantity of the set can be taken as the background noise level. This stability-based determination method can effectively avoid misjudging transient low-energy fluctuations as background noise, thereby improving the accuracy of background noise estimation.

[0082] The technical solution of the present application solves the problem of insufficient robustness in determining the background noise level in the traditional method by introducing stability analysis of the set of low-energy points. In actual application, even in the non-leakage state, the acoustic information may have transient low-energy fluctuations, which are not real background noise. If the background noise level is directly determined based on these unstable low-energy points, it may lead to inaccurate background noise estimation, which in turn affects the subsequent judgment of the leakage signal. By identifying the set of low-energy points and performing stability analysis on them, it is possible to effectively distinguish between real, continuous background noise and transient, unstable low-energy events. Only when the set of low-energy points is determined to be stable, it is used to determine the background noise level, thereby ensuring that the determined background noise level is more representative and accurate.

[0083] The above technical solution enables a more accurate and robust determination of the background noise level of residual acoustic information during leak detection for high-temperature, high-pressure steam safety valves. This stability-analysis-based background noise determination method effectively avoids interference with background noise estimation caused by transient noise or abnormally low-energy events, improving the accuracy and reliability of background noise estimation. Consequently, when subsequently adjusting the preset energy threshold for persistent acoustic information, a more reasonable target energy threshold can be obtained, significantly improving the system's ability to distinguish true leak signals from ambient noise, reducing false alarm rates, and ensuring the accuracy and reliability of leak detection.

[0084] In some embodiments, the step of performing stability analysis on the low-energy point set to obtain a stability analysis result comprises: Calculating the dispersion of energy distribution of the low-energy point set within a preset duration threshold; The stability of the low-energy point set is determined based on the discreteness of the energy distribution to obtain a stability analysis result.

[0085] The low-energy point set refers to a group of data points with low energy values ​​identified from the short-term energy of the residual acoustic information. These low-energy points are generally considered to be potential components of background noise. To accurately determine the background noise level, it is necessary to evaluate the stability of these low-energy points.

[0086] The dispersion of energy distribution is a measure of the degree of fluctuation within a set of data points. When the dispersion of the energy distribution of a set of low-energy points within a preset duration threshold is small, it indicates that these low-energy points remain relatively stable over time, with low volatility, and are more likely to represent true background noise. Conversely, if the dispersion is large, it may indicate the presence of transient interference or non-background noise components.

[0087] Furthermore, by comparing the calculated energy distribution's dispersion with a preset stability threshold, the system determines whether the low-energy point set is sufficiently stable to accurately estimate the background noise level. If the dispersion is below the preset threshold, the set is considered stable, and its energy value can be used as a reliable indicator of the background noise level.

[0088] The technical solution of the present application can quantitatively evaluate the degree of fluctuation of these low-energy points by calculating the discreteness of the energy distribution of the low-energy point set within a preset duration threshold. Precisely because background noise usually has relatively stable characteristics, while transient interference or leakage signals may cause drastic fluctuations in energy, stable background noise and unstable interference signals can be effectively distinguished by analyzing the discreteness. When the discreteness is low, it indicates that the selected low-energy point set has good stability, thereby being able to more accurately reflect the actual background noise level. In this way, it is possible to avoid misjudging transient noise or non-leakage signals as background noise, thereby improving the accuracy of background noise estimation.

[0089] Through the above-mentioned technical solution, the present application is able to quantitatively assess the stability of the low-energy point set, thereby ensuring that the determined background noise level is based on stable and reliable acoustic information. Compared to simply selecting low-energy points as background noise, the present application introduces discreteness analysis to effectively eliminate background noise estimation bias caused by transient interference or non-background noise components, making the determination of background noise more accurate. This precise background noise level estimation further improves the accuracy and reliability of subsequent leakage signal judgment, reduces the false alarm rate, and thus improves the overall performance of the high-temperature and high-pressure steam safety valve leakage detection method.

[0090] In some embodiments, the step of calculating the dispersion of energy distribution of the low-energy point set within a preset duration threshold comprises: Obtaining energy values ​​of the low-energy point set within the preset duration threshold; Based on the energy value, calculating the interquartile range of the energy value; or, based on the energy values, calculating the median absolute deviation of the energy values; The interquartile range or the median absolute deviation is used as the dispersion of the energy distribution.

[0091] Specifically, the step of calculating the dispersion of the energy distribution of the low-energy point set within a preset duration threshold may include: First, the energy values ​​of the low-energy point set within the preset duration threshold are obtained. These energy values ​​generally refer to the short-term energy values ​​of low-energy points identified from the residual acoustic information within a specific time period. These low-energy point sets are considered potential background noise samples.

[0092] Next, based on the energy values, the interquartile range (IQR) of the energy values ​​can be calculated. The IQR is a statistic that measures the degree of dispersion in data and is defined as the difference between the upper and lower quartiles. It represents the distribution of the middle 50% of the data, is insensitive to outliers, and effectively reflects the degree of dispersion of trends in a dataset.

[0093] Alternatively, the median absolute deviation of the energy values ​​can be calculated based on the energy values. The median absolute deviation is another robust statistic for measuring data dispersion. It is defined as the median of the absolute differences between each data point and the median in a dataset. Compared to the standard deviation, the median absolute deviation is more robust to outliers and can more accurately reflect the true dispersion of the data in the presence of noise or outliers.

[0094] Finally, the interquartile range or the median absolute deviation is used as the dispersion of the energy distribution. Either of these two statistics is selected as the dispersion in order to provide a robust assessment of the stability of the energy distribution of the low energy point set.

[0095] The technical solution of the present application uses the interquartile range or median absolute deviation to calculate the energy distribution dispersion of the low-energy point set, which can effectively overcome the shortcomings of traditional dispersion measurements (such as standard deviation) that are sensitive to outliers. In the actual high-temperature and high-pressure steam safety valve leakage detection environment, the background noise may have instantaneous fluctuations or occasional interference. If these outliers are included in the dispersion calculation, it may lead to an incorrect estimation of the background noise level. By using the interquartile range or median absolute deviation, these robust statistics can more accurately reflect the inherent stability of the low-energy point set, thereby ensuring that the subsequent determination of the background noise level is more accurate. This precise dispersion calculation is the key to ensuring the adaptive adjustment of the background noise level, thereby improving the accuracy of identifying persistent real leakage signals.

[0096] Through the above technical solution, this application provides a more robust and accurate method for calculating the discreteness of background noise energy distribution. This makes the estimation of background noise levels for leak detection in high-temperature, high-pressure steam safety valves more reliable in complex and changing industrial environments, effectively reducing the risk of false alarms or missed alarms caused by background noise fluctuations. As a result, the overall detection performance and reliability of the system are significantly improved, ensuring accurate identification of true leak signals.

[0097] In some embodiments, the step of determining the stability of the low-energy point set based on the discreteness of the energy distribution includes: comparing the dispersion of the energy distribution with a preset stability threshold; If the dispersion is lower than the preset stability threshold, the low energy point set is determined to be stable.

[0098] Among them, the preset stability threshold is a pre-set value used to define the upper limit at which the energy distribution of the low-energy point set is considered to be stable. This threshold can be determined through experiments, calibration or empirical values ​​based on the actual application scenario, equipment characteristics and the expected background noise recognition accuracy to provide an objective judgment standard. Specifically, when the discreteness of the energy distribution is lower than the preset stability threshold, it indicates that the energy value of the low-energy point set fluctuates less and the distribution is relatively concentrated, which indicates that within this time period, the low-energy part of the residual acoustic information is stable and reliable, and can be used as an effective basis for determining the background noise level.

[0099] The technical solution of the present application introduces a preset stability threshold and compares it with the calculated energy distribution discreteness, thereby providing a clear and quantitative standard for judging the stability of a set of low-energy points. When the discreteness is lower than the threshold, it means that within the preset duration threshold, the low-energy portion of the residual acoustic information exhibits a high degree of consistency and stability, which eliminates the interference of transient noise or abnormal fluctuations on the background noise level estimation. It is precisely because of this objective quantitative judgment that the determined background noise level is more accurate and reliable, providing a solid foundation for the subsequent adjustment of the preset energy threshold of the continuous acoustic information.

[0100] Through the above-mentioned technical solution, the present application can effectively avoid the background noise level estimation bias caused by the instability of the low-energy point set, significantly improving the accuracy and reliability of background noise level determination. As a result, in the subsequent leakage signal judgment, the energy threshold of the continuous acoustic information can be adjusted more accurately, effectively reducing the false alarm rate and the missed alarm rate, ensuring the robustness and effectiveness of high-temperature and high-pressure steam safety valve leak detection, and further enhancing the practical value and reliability of the entire detection method.

[0101] For example, assume that when analyzing the short-time energy of the residual acoustic information, a set of low-energy point sets is identified. In order to determine the stability of the set, the discreteness of its energy distribution within the preset duration threshold is first calculated. For example, if the interquartile range is used as the discreteness indicator, the interquartile range of the low-energy point set is calculated to be 0.05 units. At the same time, the system preset stability threshold is 0.1 units. At this time, the calculated discreteness 0.05 is compared with the preset stability threshold 0.1. Since 0.05 is lower than 0.1, the system will judge the low-energy point set to be stable. This means that the low-energy acoustic information within this time period fluctuates very little and can be reliably regarded as background noise. Based on this stable low-energy point set, the system can accurately determine the background noise level of the residual acoustic information and adjust the energy threshold required for subsequent leakage judgment accordingly, thereby ensuring the accuracy of leak detection.

[0102] In some embodiments, the step of determining whether there are periodic peaks in the low-frequency spectrum in the spectrum distribution information, and determining that there is no persistent real leakage signal if there are one or more periodic peaks higher than a preset threshold and the frequencies of the periodic peaks are consistent with the dripping frequency range of water droplets, and if there are no periodic peaks, determining that there is a persistent real leakage signal includes: Determining a preset frequency interval corresponding to a frequency range of water droplets from a low-frequency frequency band in the frequency spectrum distribution information; Calculating the total spectrum energy within the preset frequency interval; The sum of the spectrum energy is compared with a preset threshold. If the sum of the spectrum energy exceeds the preset threshold, it is determined that there is no persistent real leakage signal. If the sum of the spectrum energy does not exceed the preset threshold, it is determined that there is a persistent real leakage signal.

[0103] Specifically, spectral distribution information refers to the frequency-energy (or amplitude) correspondence data obtained through spectral analysis of the instantaneous energy envelope curve. The low-frequency spectrum typically refers to the frequency range below a specific frequency (e.g., 200Hz or 500Hz), as the sound of falling water droplets is typically concentrated at lower frequencies. The preset frequency interval corresponding to the falling water droplet frequency range is the frequency range where the primary acoustic energy of falling water droplets is concentrated, determined based on experience or experimental data. For example, it can be set to 5Hz to 50Hz. The spectral energy of all frequency points within this specific frequency interval is accumulated to obtain the total acoustic energy intensity within this interval, and the sum of the spectral energy within the preset frequency interval is calculated. This sum of the spectral energy can reflect whether there is significant acoustic activity within this specific frequency interval. In practical applications, a preset threshold is a reference value used to determine whether the sum of the spectral energy is large enough to indicate falling water droplets. This threshold can be calibrated and set based on the actual application scenario, ambient noise level, and the typical acoustic characteristics of falling water droplets.

[0104] The technical solution of the present application solves the problem of misjudgment or missed judgment that may exist when relying solely on periodic peaks by expanding the basis for judging water drop noise from a single "periodic peak" to the "sum of spectral energy within a specific frequency interval." Although the sound of falling water droplets may show periodicity, its spectral characteristics do not always present clear single or multiple periodic peaks. Especially in complex noise environments, the periodicity may be masked or not obvious. By calculating the sum of the spectral energy within a preset frequency interval corresponding to the frequency range of water droplets, the overall energy characteristics of the water drop noise can be captured more comprehensively. Even if its periodicity is not obvious, it can be effectively identified as long as the total energy within the frequency interval reaches a certain level. Therefore, this method can more robustly exclude non-leakage signals caused by water droplets, and avoid misjudging the sound of water droplets as real leaks.

[0105] Through the above technical solution, this application can more accurately identify and eliminate non-leakage signals caused by dripping water, significantly reducing the false alarm rate. Compared to relying solely on periodic peaks for judgment, this technical solution improves the accuracy and robustness of water droplet noise recognition by comprehensively considering the sum of spectral energy in specific low-frequency bands. This enables the high-temperature and high-pressure steam safety valve leak detection system to more reliably determine whether there is a persistent real leak signal, improving the overall performance and practical value of the system.

[0106] As a specific embodiment, when leak testing a high-temperature, high-pressure steam safety valve, the system first receives an operator command to activate ambient noise suppression mode. It then performs a first processing path to suppress the raw acoustic information collected in real time, generating target acoustic information. Subsequently, the second processing path performs a point-by-point subtraction between the raw and target acoustic information to generate residual acoustic information. The residual acoustic information is analyzed for persistent acoustic information that meets preset duration and energy thresholds. If such information exists, spectral distribution information is extracted from its instantaneous energy envelope curve. Furthermore, the system identifies a preset frequency range corresponding to the frequency range of water droplets within the low-frequency spectrum (e.g., a frequency range of 5Hz to 50Hz) within this spectral distribution information. The system then calculates the sum of the spectral energy within this preset frequency range. For example, if the calculated sum of the spectral energy is 100 and the preset threshold is 50, if the sum of the spectral energy exceeds the preset threshold, the system determines that there is no persistent true leak signal, indicating that the persistent acoustic information is likely caused by water droplets. Conversely, if the total spectral energy is 30, which does not exceed the preset threshold of 50, the system determines that a persistent true leak signal is present, as the energy characteristics of this low-frequency band do not conform to the typical pattern of dripping water. This effectively prevents the misinterpretation of dripping noise as steam leaks, thereby improving leak detection accuracy.

[0107] On the other hand, Figure 2 As shown, a high-temperature and high-pressure steam safety valve leakage detection system is exemplarily shown. The present application further proposes a high-temperature and high-pressure steam safety valve leakage detection system 100, which includes: The receiving module 10 is configured to receive an instruction from an operator to enable the ambient noise suppression mode; a first path processing module 20 for performing a first processing path suppression process on the raw acoustic information collected in real time in response to the ambient noise suppression mode activation instruction, generating target acoustic information, and reporting the valve status of the high-temperature and high-pressure steam safety valve based on the target acoustic information; a second path processing module 30, configured to subtract the original acoustic information from the target acoustic information point by point through a second processing path to generate residual acoustic information; An independent analysis module 40 is configured to analyze the residual acoustic information to determine whether there is persistent acoustic information that meets a preset duration threshold and a preset energy threshold, so as to determine whether there is a persistent true leakage signal; The early warning recording module 50 is configured to trigger an early warning record if it is determined that the persistent true leakage signal exists.

[0108] The high-temperature and high-pressure steam safety valve leakage detection system disclosed in this application may include the following modules: First, the system includes a receiving module for receiving an operator-issued instruction to activate the ambient noise suppression mode. The specific method for receiving the operator-issued instruction to activate the ambient noise suppression mode has been described in the above embodiments and will not be repeated here. It should be emphasized that the receiving module in this application can be configured in a variety of forms. For example, it can be a physical interface, such as a USB port, Ethernet port, or wireless communication module (such as a Wi-Fi, Bluetooth, or cellular network module), for receiving instructions from an external device or network. As another embodiment, the receiving module can also be a software component within the system, responsible for monitoring specific events or message queues to capture instructions issued by the operator through the user interface. In some implementations, the receiving module may also include an analog or digital input circuit for receiving electrical signals from a physical button or switch. These implementations may involve considerations regarding data transmission rate, anti-interference capabilities, or compatibility.

[0109] Secondly, the system includes a first-path processing module, configured to, in response to the aforementioned instruction to activate the ambient noise suppression mode, perform a first-path suppression process on the raw acoustic information collected in real time, generate target acoustic information, and report the valve status of the high-temperature, high-pressure steam safety valve based on the target acoustic information. The specific methods for performing the first-path suppression process on the raw acoustic information collected in real time, generating target acoustic information, and reporting the valve status of the high-temperature, high-pressure steam safety valve based on the target acoustic information have been described in the aforementioned embodiments and will not be repeated here. It should be emphasized that the first-path processing module in this application can be implemented as a dedicated hardware circuit, such as a digital signal processor (DSP) or a field-programmable gate array (FPGA), to achieve efficient real-time acoustic information processing. Alternatively, it can be a software program running on a general-purpose processor (CPU) that performs noise suppression by executing a pre-defined algorithm (such as spectral subtraction, adaptive filtering, or wavelet denoising). In some hybrid implementations, this module may include a front-end analog signal processing unit and a back-end digital processing unit. These implementations may involve trade-offs in processing speed, power consumption, or algorithmic flexibility.

[0110] Furthermore, the system includes a second path processing module for subtracting the above-mentioned original acoustic information from the above-mentioned target acoustic information point by point through the second processing path to generate residual acoustic information. The specific method of subtracting the original acoustic information from the target acoustic information point by point through the second processing path to generate residual acoustic information has been recorded in the above embodiment and will not be repeated here. It should be emphasized that the second path processing module in the present application can be implemented as a hardware logic circuit, such as a subtractor or a digital comparator, for performing point-by-point subtraction operations on the input digital acoustic data stream. As another embodiment, it can also be a software module running on a processor, which implements synchronous reading and point-by-point subtraction operations on two acoustic information sequences through programming. In some high-performance systems, the module can use a parallel processing architecture to accelerate subtraction operations. These implementations may be challenging in terms of synchronization accuracy, computing resource usage or data throughput.

[0111] Furthermore, the system includes an independent analysis module for analyzing whether there is persistent acoustic information in the above-mentioned residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal. The above-mentioned embodiment has recorded the specific method of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal, which will not be repeated here. It should be emphasized that the independent analysis module in the present application can be implemented as firmware on an embedded processor, which is responsible for executing complex signal processing and pattern recognition algorithms to identify specific patterns in the residual acoustic information. As another embodiment, it can also be a software service running on a server or cloud computing platform, which performs offline or near real-time analysis by receiving a residual acoustic information data stream. In some edge computing scenarios, the module can be deployed on a local device to reduce data transmission delay. These implementations may differ in algorithm complexity, real-time performance, or resource consumption.

[0112] Finally, the system includes an early warning recording module, which is used to trigger an early warning record if it is determined that the above-mentioned continuous real leakage signal exists. The specific method of triggering the early warning record if it is determined that the continuous real leakage signal exists has been recorded in the above embodiment and will not be repeated here. It should be emphasized that the early warning recording module in this application can be implemented as a hardware unit with a storage function, such as a non-volatile memory (such as flash memory or EEPROM), which is used to record the early warning event log. As another embodiment, it can also be part of a software application, responsible for writing the early warning information to a local database or sending it to a remote monitoring system via a network. In some integrated systems, the module can also control an audible and visual alarm or send SMS / email notifications. These implementation methods may have limitations in data storage capacity, communication reliability or alarm response speed.

[0113] The core innovation of the high-temperature and high-pressure steam safety valve leakage detection system disclosed in this application lies in the realization of a dual-path acoustic information processing mechanism through modular design, combined with the analysis of residual acoustic information, to solve the dilemma of false alarms and missed alarms coexisting in traditional systems in complex industrial environments, especially in the presence of water droplet noise interference and environmental noise suppression mode. Compared with the systems in the prior art that rely only on a single noise suppression path for leak detection, the system of this application has the following advantages: when the environmental noise suppression mode is enabled, the traditional system often indiscriminately filters out all signals identified as noise, which may include weak but real continuous steam leakage signals, resulting in missed alarms. The system of this application uses the second path processing module to perform point-by-point subtraction of the original acoustic information from the target acoustic information after suppression processing by the first path processing module, thereby generating residual acoustic information. This innovative design allows potential leakage signals that are suppressed or weakened in the first path processing module to be "restored" or "highlighted." By using an independent analysis module to determine the duration threshold and energy threshold of the above-mentioned residual acoustic information, the system of the present application can effectively identify real leakage signals with continuous characteristics and distinguish them from transient interference (such as the sound of water drops). The sound of water drops usually appears as a short-term pulse, while steam leaks are continuous. Therefore, even in the environmental noise suppression mode, the system of the present application can accurately capture weak continuous leakage signals and avoid missed reports. At the same time, by analyzing the characteristics of the residual signal, false alarms caused by non-leakage noises such as water drops are also effectively avoided. In summary, the system of the present application achieves accurate and reliable detection of high-temperature and high-pressure steam safety valve leakage in complex noise environments through modular dual-path processing and residual acoustic information analysis, significantly improving the anti-interference ability and recognition accuracy of the detection system, thereby effectively ensuring the safety of industrial production and the operational stability of equipment, and reflecting significant progress.

[0114] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for detecting leakage of a high-temperature and high-pressure steam safety valve, characterized in that: include: receiving an ambient noise suppression mode activation instruction from an operator; In response to the instruction to enable the ambient noise suppression mode, performing suppression processing of the first processing path on the raw acoustic information collected in real time to generate target acoustic information, and reporting the valve status of the high-temperature and high-pressure steam safety valve based on the target acoustic information; subtracting the original acoustic information from the target acoustic information point by point through a second processing path to generate residual acoustic information; Analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold, so as to determine whether there is a persistent real leakage signal; If it is determined that the persistent real leakage signal exists, an early warning record is triggered.

2. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 1, characterized in that: The step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal further includes: If there is continuous acoustic information that meets a preset duration threshold and a preset energy threshold, extracting an instantaneous energy envelope curve from the continuous acoustic information; Performing spectrum analysis on the instantaneous energy envelope curve to obtain spectrum distribution information; Determine whether there are periodic peaks in the low-frequency band spectrum in the spectrum distribution information. If there are one or more periodic peaks higher than a preset threshold, and the frequency of the periodic peaks is consistent with the dripping frequency range of water droplets, it is determined that there is no persistent real leakage signal. If there are no periodic peaks, it is determined that there is a persistent real leakage signal.

3. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 1, characterized in that: The step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold to determine whether there is a persistent real leakage signal further includes: If there is continuous acoustic information that meets a preset duration threshold and a preset energy threshold, accurately aligning the energy change envelope of the continuous acoustic information with the amplitude change envelope of the vibration signal collected by one or more mechanical vibration sensors preset in the steam pipe system in time; Within a preset analysis period, the correlation coefficient between the energy change envelope of the continuous acoustic information and the amplitude change envelope of the vibration signal is calculated; if the correlation coefficient is higher than the preset correlation coefficient threshold, it is determined that there is no continuous real leakage signal; if the correlation coefficient is lower than the preset correlation coefficient threshold, it is determined that there is a continuous real leakage signal.

4. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 1, characterized in that: The step of analyzing whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold includes: Calculating the short-time energy of the residual acoustic information; Counting the distribution of the short-term energy within a preset duration threshold; determining a background noise level of the residual acoustic information based on a preset low-energy portion of the distribution of the short-term energy; Adjusting the preset energy threshold of the continuous acoustic information according to the background noise level to obtain a target energy threshold; It is determined whether the short-term energy is within a preset duration threshold and is maintained above the target energy threshold.

5. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 4, characterized in that: The step of determining the background noise level of the residual acoustic information based on the preset low-energy portion of the distribution of the short-time energy comprises: identifying a set of low-energy points from the short-time energies; Performing stability analysis on the low-energy point set to obtain a stability analysis result; Based on the stability analysis result, a background noise level of the residual acoustic information is determined.

6. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 5, characterized in that: The step of performing stability analysis on the low-energy point set to obtain a stability analysis result comprises: Calculating the dispersion of energy distribution of the low-energy point set within a preset duration threshold; The stability of the low-energy point set is determined based on the discreteness of the energy distribution to obtain a stability analysis result.

7. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 6, characterized in that: The step of calculating the dispersion of the energy distribution of the low-energy point set within a preset duration threshold comprises: Obtaining energy values ​​of the low-energy point set within the preset duration threshold; Based on the energy value, calculating the interquartile range of the energy value; or, based on the energy values, calculating the median absolute deviation of the energy values; The interquartile range or the median absolute deviation is used as the dispersion of the energy distribution.

8. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 6, characterized in that: The step of judging the stability of the low-energy point set according to the discreteness of the energy distribution includes: comparing the dispersion of the energy distribution with a preset stability threshold; If the dispersion is lower than the preset stability threshold, the low energy point set is determined to be stable.

9. The high-temperature and high-pressure steam safety valve leakage detection method according to claim 2, characterized in that: The step of determining whether there are periodic peaks in the low-frequency spectrum in the spectrum distribution information, and determining that there is no persistent real leakage signal if there are one or more periodic peaks higher than a preset threshold and the frequencies of the periodic peaks are consistent with the dripping frequency range of water droplets, and determining that there is no persistent real leakage signal if there are no periodic peaks, includes: Determining a preset frequency interval corresponding to a frequency range of water droplets from a low-frequency frequency band in the frequency spectrum distribution information; Calculating the total spectrum energy within the preset frequency interval; The sum of the spectrum energy is compared with a preset threshold. If the sum of the spectrum energy exceeds the preset threshold, it is determined that there is no persistent real leakage signal. If the sum of the spectrum energy does not exceed the preset threshold, it is determined that there is a persistent real leakage signal.

10. A high-temperature and high-pressure steam safety valve leakage detection system, characterized in that: The system includes: A receiving module, configured to receive an instruction from an operator to enable an ambient noise suppression mode; a first path processing module, configured to, in response to the ambient noise suppression mode activation instruction, perform a first processing path suppression process on the raw acoustic information collected in real time to generate target acoustic information, and report the valve status of the high-temperature and high-pressure steam safety valve based on the target acoustic information; a second path processing module, configured to subtract the original acoustic information from the target acoustic information point by point through a second processing path to generate residual acoustic information; An independent analysis module is used to analyze whether there is persistent acoustic information in the residual acoustic information that meets a preset duration threshold and a preset energy threshold, so as to determine whether there is a persistent real leakage signal; The early warning recording module is used to trigger the early warning recording if it is determined that the persistent real leakage signal exists.

Images