A safety monitoring method and system for a gas pressure reducer

Abstract

The application discloses a kind of safety monitoring method and system of gas pressure reducer, belong to mechanical equipment state monitoring technical field, it includes obtaining the outlet pressure, flow, valve body temperature and vibration signal of gas pressure reducer, obtains multi-source monitoring data;Excitation is applied to gas pressure reducer, and the change of multi-source monitoring data during excitation is synchronously collected, and excitation response data is generated;Pressure recovery time, vibration spectrum change and temperature change gradient are extracted from excitation response data, combined into dynamic feature set;Dynamic feature set is analyzed to identify the response mode of the health state of internal mechanical components;Based on response mode, the health state of gas pressure reducer is quantitatively evaluated, and health index is generated.The means of active excitation and coupling analysis of multi-source response data can identify early failure mode of internal mechanical components and quantitatively evaluate its health state.

Description

Technical Field

[0001] This invention relates to the field of mechanical equipment condition monitoring technology, and in particular to a safety monitoring method and system for a gas pressure regulator. Background Technology

[0002] A gas pressure regulator, also known as a pressure regulating valve, is a key fluid control component that reduces the pressure of gas at the inlet to a desired, stable outlet pressure. It is widely used in petrochemical, energy, aerospace, and various industrial gas pipeline systems. Its operational stability and reliability directly affect the safety of downstream equipment and the smooth operation of the entire process, making it one of the core components ensuring production safety.

[0003] Currently, safety monitoring methods for gas pressure regulators mainly rely on periodic manual inspections and monitoring of a single parameter, the outlet pressure. Manual inspections involve operators making rough judgments by listening, looking, and touching, which is highly subjective and cannot achieve continuous monitoring. Online monitoring systems based on pressure sensors typically only focus on whether the outlet pressure fluctuates within a set normal range, triggering an alarm only when a significant pressure anomaly occurs—a reactive monitoring approach. In some cases, offline disassembly and maintenance are also used to assess the internal condition.

[0004] However, the aforementioned existing technologies have significant technical drawbacks. Monitoring methods based on a single outlet pressure parameter have extremely low sensitivity in detecting early performance degradation of internal mechanical components, such as elastic fatigue of the pressure regulating spring and minor wear of the valve core and seat. In these early stages of failure, the pressure reducer can often barely maintain the apparent stability of the outlet pressure, thus masking potential safety hazards. By the time significant anomalies in the pressure parameters appear, internal damage is often already quite severe. Furthermore, a single pressure signal cannot provide sufficient fault diagnosis information, making it difficult to distinguish the specific internal causes of pressure instability, leading to blind spots in maintenance work. Offline maintenance, on the other hand, interrupts production, is costly, and cannot reflect the dynamic performance of the equipment under actual operating conditions. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a safety monitoring method and system for gas pressure regulators. By actively applying operating condition excitation and performing coupled analysis of multi-source response data, it is possible to accurately identify early failure modes of internal mechanical components and quantitatively assess their health status.

[0006] The above objectives can be achieved through the following approach:

[0007] A safety monitoring method for a gas pressure regulator includes acquiring outlet pressure, flow rate, valve body temperature, and vibration signals of the gas pressure regulator to obtain multi-source monitoring data; applying operating condition excitation to the gas pressure regulator and simultaneously collecting changes in the multi-source monitoring data during the excitation period to generate excitation response data; extracting pressure recovery time, vibration spectrum changes, and temperature change gradients from the excitation response data and combining them into a dynamic feature set; performing coupled response analysis on the dynamic feature set to identify response patterns characterizing the health status of internal mechanical components; and quantitatively evaluating the health status of the gas pressure regulator based on the response patterns to generate a health index.

[0008] Optionally, obtaining the multi-source monitoring data includes: acquiring the outlet pressure through a pressure sensor, acquiring the flow rate through a flow sensor, acquiring the valve body temperature through a temperature sensor, and acquiring the vibration signal through a vibration sensor; synchronizing and aligning the outlet pressure, flow rate, valve body temperature, and vibration signal in time, and combining them into multi-source monitoring data.

[0009] Optionally, generating the excitation response data includes: controlling a miniature solenoid valve located downstream of the gas pressure reducer to generate a flow disturbance as an operating condition excitation; during the application of the operating condition excitation, collecting changes in multi-source monitoring data at a preset sampling rate to obtain the excitation response data.

[0010] Optionally, the combination into a dynamic feature set includes: calculating the time required for the outlet pressure to recover to a preset stable state range from the excitation end time, to obtain the pressure recovery time; calculating the energy change of the vibration signal during the excitation period relative to before the excitation through spectrum analysis, to obtain the vibration spectrum change; calculating the rate of change of the valve body temperature during the excitation period, to obtain the temperature change gradient; and combining the pressure recovery time, the vibration spectrum change, and the temperature change gradient into a dynamic feature set.

[0011] Optionally, the coupled response analysis of the dynamic feature set to identify response patterns characterizing the health status of internal mechanical components includes: analyzing the correlation between the pressure recovery time and the vibration spectrum change to distinguish between changes in system damping characteristics and abnormal mechanical impacts; and matching the dynamic feature set with a preset pattern library characterizing mechanical health status based on the correlation to identify the response patterns.

[0012] Optionally, the response modes include a health mode, a spring fatigue mode, and a valve core wear mode; wherein, the identification of the spring fatigue mode is based on the correlation between the prolonged pressure recovery time and the occurrence of oscillations, and the continuous abnormality of the vibration spectrum change in the low-frequency band; the identification of the valve core wear mode is based on the occurrence of pressure overshoot during the pressure recovery time, and the correlation between the occurrence of new spectral peaks in the vibration spectrum change in the high-frequency band.

[0013] Optionally, the step of quantitatively evaluating the health status of the gas pressure regulator and generating a health index based on the response pattern includes: mapping the identified response pattern to a preset health index value range to generate an initial index; calculating the health of the mechanical component based on the degree of deviation of the dynamic features in the response pattern to obtain the component health; and fusing the component health and the initial index to obtain a health index.

[0014] Optionally, calculating the health of the mechanical component based on the deviation of the dynamic characteristics in the response mode to obtain the component health includes: when the response mode is a spring fatigue mode, calculating the spring health based on the oscillation degree of the pressure recovery time and the abnormality of the vibration spectrum change in the low-frequency band; when the response mode is a valve core wear mode, calculating the valve core seal health based on the overshoot amplitude of the pressure recovery time and the abnormality of the vibration spectrum change in the high-frequency band; the component health includes the spring health and the valve core seal health.

[0015] Optionally, the method further includes: performing time-series tracking on the health index to generate a health index change rate; generating an early warning signal when the health index is lower than a preset safety threshold or the health index change rate is less than zero; and generating an early warning report based on the early warning signal and the corresponding component health.

[0016] Based on the same inventive concept, this invention also provides a safety monitoring system for a gas pressure regulator. The system includes: a multi-source sensing module for acquiring outlet pressure, flow rate, valve body temperature, and vibration signals of the gas pressure regulator to obtain multi-source monitoring data; an operating condition excitation module for applying operating condition excitation to the gas pressure regulator and simultaneously collecting changes in the multi-source monitoring data during the excitation period to generate excitation response data; a dynamic feature extraction module for extracting pressure recovery time, vibration spectrum changes, and temperature change gradients from the excitation response data and combining them into a dynamic feature set; a coupling analysis module for performing coupling response analysis on the dynamic feature set to identify response patterns characterizing the health status of internal mechanical components; and a health assessment module for quantitatively assessing the health status of the gas pressure regulator based on the response patterns and generating a health index.

[0017] Compared with the prior art, the present invention has the following advantages:

[0018] 1. This invention, by actively applying operating condition excitation and simultaneously collecting multi-source monitoring data, can effectively stimulate and capture early fault characteristics that are difficult to manifest when the equipment is running in a steady state. This active detection method amplifies weak performance degradation information, which improves the sensitivity and timeliness of fault detection compared with traditional passive monitoring methods, thereby achieving early warning of potential safety risks.

[0019] 2. This invention establishes a direct mapping relationship from multidimensional data features to specific internal mechanical component failure modes by performing coupled response analysis on dynamic characteristics such as pressure recovery time and vibration spectrum changes. This method can accurately distinguish failures with different physical causes such as spring fatigue and valve core wear, achieving accurate fault location and diagnosis. It overcomes the technical bottleneck of traditional single-parameter monitoring methods that are difficult to identify fault types and improves the accuracy of diagnosis.

[0020] 3. The present invention ultimately generates a quantitative health index, which transforms the complex and abstract health status of equipment into an intuitive and traceable value. This quantitative assessment result can not only reflect the type of failure, but also measure its severity, providing a scientific basis for the analysis of equipment health status trends, prediction of remaining service life, and the formulation of condition-based precision maintenance strategies, thereby improving the intelligence and refinement of equipment management.

[0021] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating a safety monitoring method for a gas pressure regulator according to an embodiment of the present invention.

[0024] Figure 2 This is a schematic diagram of the multi-source monitoring data response curve according to an embodiment of the present invention.

[0025] Figure 3 This is a schematic diagram illustrating the tracking and early warning of the evolution of the health index of the gas pressure regulator over time according to an embodiment of the present invention.

[0026] Figure 4This is a schematic diagram of the structure of a safety monitoring system for a gas pressure regulator according to an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Reference Figure 1 One embodiment of the present invention proposes a safety monitoring method for a gas pressure regulator. By actively applying operating condition excitation and performing coupled analysis of multi-source response data, it is possible to accurately identify early failure modes of internal mechanical components and quantitatively assess their health status.

[0029] The method described in this embodiment specifically includes:

[0030] The outlet pressure, flow rate, valve body temperature, and vibration signal of the gas pressure regulator are acquired to obtain multi-source monitoring data.

[0031] Apply operating condition excitation to the gas pressure regulator and simultaneously collect the changes of the multi-source monitoring data during the excitation period to generate excitation response data;

[0032] The pressure recovery time, vibration spectrum change, and temperature change gradient are extracted from the excitation response data and combined into a dynamic feature set.

[0033] Coupled response analysis is performed on the dynamic feature set to identify response patterns that characterize the health status of internal mechanical components;

[0034] Based on the response pattern, the health status of the gas pressure regulator is quantitatively assessed, and a health index is generated.

[0035] This invention transforms passive monitoring into active detection. By applying controllable operating condition stimuli, it enhances the ability to detect early, minor faults, solving the problem that traditional methods struggle to effectively capture early degradation information during normal equipment operation. Secondly, the fusion and coupled analysis of multi-source information improves the accuracy of fault diagnosis, enabling the differentiation of similar symptoms caused by different physical factors, achieving precise location from "whether there is a fault" to "where the fault is and what type of fault it is." Finally, by generating a quantified health index, the abstract equipment health status is transformed into an intuitive and traceable value, providing scientific and reliable data support for the equipment's full lifecycle health management, predictive maintenance, and repair decisions, thereby improving the operational reliability and safety of the gas pressure regulator.

[0036] Optionally, the obtained multi-source monitoring data includes:

[0037] The outlet pressure is collected by a pressure sensor, the flow rate is collected by a flow sensor, the valve body temperature is collected by a temperature sensor, and the vibration signal is collected by a vibration sensor.

[0038] The outlet pressure, flow rate, valve body temperature, and vibration signals are synchronized and aligned in time, and combined into multi-source monitoring data.

[0039] Specifically, a pressure sensor is installed on the outlet pipeline of the gas pressure regulator to collect the outlet pressure signal in real time, and a flow sensor is installed to collect the corresponding flow signal. A temperature sensor is installed on the surface of the pressure regulator valve body, especially near the internal valve core and pressure regulating mechanism, to monitor the valve body temperature. In addition, a vibration sensor, typically a high-frequency accelerometer, is installed on the rigid structural part of the valve body shell to collect vibration signals generated during equipment operation. All sensors are connected to a data acquisition system with multi-channel synchronous acquisition capabilities. To achieve time synchronization alignment of the multi-source monitoring data, the system uses a unified clock source to synchronously sample signals from all channels. During the acquisition process, the system performs analog-to-digital conversion on the analog signals from the pressure sensor, flow sensor, temperature sensor, and vibration sensor at a preset sampling frequency, and assigns a precise, unified timestamp to each sampling point. This operation ensures that at any given time, the obtained outlet pressure, flow rate, valve body temperature, and vibration signal values ​​accurately correspond to the physical state at the same moment. Finally, these time-synchronized time-series data containing multiple physical quantities are combined to form a multi-dimensional data matrix, i.e., multi-source monitoring data. This data can be represented as: ,

[0040] in The data is from multiple sources at time t. Due to export pressure, For traffic, For valve body temperature, This is a vibration signal. For example... Figure 2 As shown in the figure, this illustrates a typical operating condition excitation and the response process of a healthy pressure regulator. At t=0.5s, a brief flow disturbance is generated by controlling the downstream solenoid valve; this is the operating condition excitation. In response, the outlet pressure experiences a brief drop before quickly and smoothly recovering to the set value of 0.5 MPa. Simultaneously, the vibration signal exhibits only small random fluctuations during the excitation period, and the valve body temperature changes gradually. This set of curves constitutes a "response fingerprint" of the healthy state, providing a benchmark for subsequent fault diagnosis.

[0041] Optionally, the generated stimulus response data includes:

[0042] The flow disturbance generated by the miniature solenoid valve located downstream of the gas pressure regulator is used as the operating condition excitation.

[0043] During the application of the aforementioned operating condition excitation, changes in multi-source monitoring data are collected at a preset sampling rate to obtain excitation response data.

[0044] Specifically, a miniature solenoid valve needs to be connected in series in the downstream pipeline of the gas pressure regulator. This solenoid valve is precisely driven by an independent control unit. The application of the operating condition excitation is based on the instantaneous state switching of this miniature solenoid valve. When the gas pressure regulator is in a stable operating state, i.e., its outlet pressure is maintained near the set value, the control unit sends a preset pulse control signal to the miniature solenoid valve. This signal drives the solenoid valve to complete a rapid opening and closing action in a very short time. This action will instantly change the downstream gas load, causing a brief but violent flow disturbance. This artificially introduced, controllable flow disturbance is the operating condition excitation described in this method. Throughout the entire period of applying the operating condition excitation, from the short period before the miniature solenoid valve begins to act, to its completion, and then to the dynamic adjustment inside the gas pressure regulator and its eventual return to a stable operating state, a pre-deployed multi-source sensing module continuously operates at a preset sampling rate. This sampling rate must be set according to the Nyquist sampling theorem to ensure that it is high enough to capture the transient changes in pressure, flow, temperature, and vibration signals caused by the operating condition excitation without distortion. The data acquisition system records the complete sequence of multi-source monitoring data collected synchronously within this specific time window, forming a dataset containing information about the entire process of the system from steady state to transient state and then back to steady state. This dataset is the excitation response data.

[0045] Optionally, the combination into a dynamic feature set includes:

[0046] The time required for the outlet pressure to recover from the excitation end time to a preset steady state range is calculated to obtain the pressure recovery time;

[0047] The energy change of the vibration signal during excitation relative to before excitation is calculated by spectrum analysis to obtain the vibration spectrum change.

[0048] Calculate the rate of change of the valve body temperature during the excitation period to obtain the temperature change gradient;

[0049] The pressure recovery time, the vibration spectrum change, and the temperature change gradient are combined into a dynamic feature set.

[0050] Specifically, the pressure recovery time is calculated first. This calculation utilizes the outlet pressure time series from the excitation response data. First, the precise moment when the excitation ends under the operating condition is determined, denoted as... Simultaneously, based on the normal operating set pressure of the gas pressure regulator, a stable range is preset, which is typically a small percentage fluctuation range of the set pressure. Then, from... Starting from point A, search backwards in the pressure time series to find the moment when the pressure value first re-enters and remains within the aforementioned steady-state range, denoted as point B. The pressure recovery time is calculated as the time difference between these two points in time. The formula is: ,

[0051] in, For pressure recovery time, The moment when the pressure returns to a stable state. This refers to the excitation end time. Both of these time values ​​can be directly obtained from the timestamps of the excitation response data.

[0052] Next, the vibration spectrum variation is calculated. This calculation is performed on the vibration signal in the excitation response data. First, the vibration signal is divided into two time periods: a stable operating signal before excitation and a transient signal during the excitation process. A Fast Fourier Transform (FFT) is then performed on the signals for each of these two time periods to obtain their respective power spectral density functions, denoted as... and By integrating the power spectral density function over the entire frequency domain, the total vibrational energy before and during excitation can be calculated separately. The change in the vibrational spectrum is defined as the difference between these two energy values, used to quantify the additional vibrational energy induced by the excitation. ,

[0053] in, For the change of vibration spectrum, For frequency, and These are the power spectral density functions of the vibration signal before and during excitation, respectively.

[0054] Then, the temperature gradient is calculated. This calculation is based on the valve body temperature time series in the excitation response data. The rate of change of temperature over time is calculated by processing the temperature data during the excitation period. A direct calculation method is to take the start time of the excitation period. and end time The corresponding temperature value is used to calculate its linear slope, which is then used as an approximation of the temperature change gradient. ,

[0055] in, For temperature gradient, and These are the valve body temperatures at the end and start of the excitation, respectively.

[0056] Finally, the pressure recovery time calculated through the above three steps will be used. Vibration spectrum changes and temperature gradient These three scalar values ​​are combined to form a three-dimensional feature vector, which is the dynamic feature set.

[0057] Optionally, the coupled response analysis of the dynamic feature set to identify response patterns characterizing the health status of internal mechanical components includes:

[0058] Analyze the correlation between the pressure recovery time and the vibration spectrum change to distinguish between changes in system damping characteristics and abnormal mechanical shocks;

[0059] Based on the correlation, the dynamic feature set is matched with a preset pattern library representing the health status of machinery to identify the response pattern.

[0060] Specifically, this method performs coupled response analysis on the dynamic feature set extracted in the previous step. Its core lies in uncovering the intrinsic correlations between different physical quantity features, rather than evaluating each feature in isolation. This analysis aims to identify unique combination patterns, i.e., response patterns, from the multidimensional feature space that can uniquely point to the health status of a specific internal mechanical component. The first step of the analysis is to delve into the correlation between pressure recovery time and changes in the vibration spectrum. This correlation analysis is not a simple linear correlation calculation, but rather pattern identification based on physical mechanisms. For example, changes in system damping characteristics, such as fatigue of the regulating spring leading to a decrease in its stiffness, typically manifest as a significant extension of the pressure recovery time. The adjustment process may be accompanied by low-frequency oscillations, which can cause energy concentration in the low-frequency range of the vibration spectrum. Abnormal mechanical shocks, such as impacts between the valve core and seat due to wear or jamming, while also affecting the pressure recovery process, are more significantly characterized by transient, broadband energy surges or new characteristic spectral peaks in the high-frequency range of the vibration spectrum. By constructing an analytical model, this model can distinguish between two completely different physical processes based on the correspondence between the changes in pressure recovery time (such as prolongation, oscillation, and overshoot) and the energy distribution characteristics of vibration spectrum changes in different frequency bands. Based on the above correlation analysis, the second step is pattern matching. This method requires the pre-establishment of a pattern library characterizing the mechanical health state. This pattern library is formed by conducting operating condition excitation experiments on the gas pressure regulator under different health states, such as brand new state, known spring fatigue state, and known valve core wear state, extracting its dynamic feature set, and establishing a standard feature template for each type of health state. The matching process compares the dynamically acquired feature set in real-time monitoring, especially the coupling characteristics of pressure recovery time and vibration spectrum changes, with each standard template stored in the pattern library. The comparison can employ algorithms from the field of pattern recognition, such as calculating the distance or similarity between the feature vector and the center of each template. When the real-time feature has the highest matching degree with a template in the library and exceeds a preset confidence threshold, the health state represented by that template is identified as the response mode of the current gas pressure regulator.

[0061] Optionally, the response modes include a health mode, a spring fatigue mode, and a valve core wear mode;

[0062] The identification of the spring fatigue mode is based on the correlation between the prolonged pressure recovery time and the occurrence of oscillations, and the continuous abnormality of the vibration spectrum change in the low-frequency range.

[0063] The identification of the valve core wear mode is based on the correlation between the pressure overshoot that occurs during the pressure recovery time and the appearance of a new spectral peak in the high-frequency band of the vibration spectrum change.

[0064] Specifically, this method defines the response modes identified in coupled response analysis, clarifying the healthy mode, spring fatigue mode, and valve core wear mode, and elucidating the identification criteria for the latter two failure modes. These criteria are based on a deep understanding of the specific coupling relationships between various features in the dynamic feature set. First, for the identification of the healthy mode, the criteria are a short and stable pressure recovery time, and vibration spectrum changes and temperature gradients both within a baseline range obtained from a large amount of normal operation data. This indicates that the internal mechanical components of the pressure reducer respond quickly, the dynamic process is smooth, and there is no abnormal energy release. Second, the identification of the spring fatigue overrun mode is based on the simultaneous occurrence of the following two coupled characteristics. First, the pressure recovery time shows a significantly prolonged trend, and during the recovery process, the outlet pressure curve no longer converges smoothly to the set value, but exhibits obvious and continuous low-frequency oscillations. This oscillating behavior is due to the decrease in the elastic coefficient, i.e., stiffness, of the regulating spring due to fatigue, which worsens the damping characteristics of the entire regulating system and reduces its stability. Second, corresponding to the pressure oscillations, vibration spectrum analysis showed a sustained and abnormal energy concentration or new spectral peaks in the low-frequency range, which is usually related to the natural frequency of the spring-diaphragm system. This abnormal enhancement of low-frequency vibration confirmed the instability of the pressure regulation system itself, rather than being caused by external random disturbances. Therefore, when the pressure recovery time is prolonged and accompanied by oscillations, and there is a strong correlation between these two phenomena and the sustained abnormality of the vibration spectrum in the low-frequency range, the system identifies the current response mode as a spring fatigue mode. Third, the identification of the valve core wear mode is also based on a specific set of coupling characteristics. First, the pressure recovery time exhibits a pressure overshoot phenomenon in its morphology. That is, after the excitation of the operating condition ends, the outlet pressure does not stabilize directly, but first rises rapidly beyond the upper limit of the preset stable state range, forming a significant pressure peak, and then falls back to the stable value. This overshoot phenomenon is due to the gap created by wear on the sealing surface between the valve core and the valve seat, resulting in the inability to achieve rapid and effective flow control at the moment the valve closes. Second, vibration spectrum analysis revealed new, discrete peaks in the high-frequency band. These high-frequency peaks are typically associated with high-frequency acoustic emissions generated by intermetallic impact, friction, or gas leakage. These new peaks directly reflect abnormal contact caused by wear between the valve core and seat during operation, or impact vibrations caused by high-speed gas leakage. Therefore, when pressure overshoot occurs during pressure recovery, and this corresponds clearly to the appearance of new peaks in the high-frequency vibration spectrum, the system identifies the response mode as the valve core wear mode.

[0065] Optionally, the step of quantitatively assessing the health status of the gas pressure regulator based on the response mode and generating a health index includes:

[0066] The identified response patterns are mapped to a preset health index range to generate an initial index.

[0067] Based on the degree of deviation of the dynamic characteristics in the response mode, the health of the mechanical component is calculated to obtain the component health.

[0068] The health status of the component and the initial index are combined to obtain the health status index.

[0069] Specifically, the first step is to map the identified response patterns to a preset health index range, generating an initial index. This process requires establishing a health index evaluation system, such as a scoring system from 0 to 100, where 100 represents complete health and 0 represents complete failure. Then, a fixed numerical range representing the basic severity of each predefined response pattern is assigned. For example, a healthy pattern might correspond to the [90-100] range, a spring fatigue pattern to the [50-70] range, and a valve core wear pattern to the [30-60] range. When the system identifies the current response pattern as, for example, the spring fatigue pattern, a baseline value, such as the median of 60, is taken from its corresponding [50-70] range as the initial index. This initial index provides a macroscopic, categorical assessment of the equipment's health status. The second step involves refining the health of mechanical components based on the specific deviation of dynamic characteristics within the response pattern, obtaining the component health score. This step aims to quantify the severity of the failure. Guided by the identified response pattern, the system focuses on the dynamic characteristics most relevant to that pattern. For example, if the system identifies a spring fatigue mode, it will focus on analyzing the oscillation degree of the pressure recovery time and the abnormal energy of the vibration spectrum in the low-frequency range; if it identifies a valve core wear mode, it will focus on analyzing the overshoot amplitude of the pressure recovery time and the energy of the new peak in the vibration spectrum in the high-frequency range. By comparing these deviations, such as the variance of the oscillation, the peak value of the overshoot, and the energy of the abnormal peak, with the baseline values ​​under healthy conditions, and using a preset degradation model or scoring function, a specific score is calculated to characterize the health status of the specific component; this is the component health score. The third step is to fuse the component health score with the initial index to obtain the final health score index. The fusion process aims to combine macroscopic classification assessment with microscopic quantitative assessment to generate a more accurate and comprehensive overall health score. An effective fusion strategy is to use the health score index range determined in the first step and use the component health score calculated in the second step as the basis for precise positioning within that range. The calculation formula can be expressed as: ,

[0070] in, For the final health index, and These are the lower and upper limits of the health index value range corresponding to the current response mode, respectively. It is the calculated component health status. This is the theoretical maximum value of component health, i.e., the value under fully healthy conditions. Using this formula, the initial index determines the basis of health, and the component health is then fine-tuned within this basis to obtain a comprehensive health index that fully reflects the type and severity of the fault.

[0071] Optionally, calculating the health of the mechanical component based on the degree of deviation of the dynamic characteristics in the response mode to obtain the component health includes:

[0072] When the response mode is spring fatigue mode, the spring health is calculated based on the degree of oscillation of the pressure recovery time and the degree of abnormality of the vibration spectrum change in the low frequency band.

[0073] When the response mode is the valve core wear mode, the valve core sealing health is calculated based on the overshoot amplitude of the pressure recovery time and the degree of abnormality of the vibration spectrum change in the high frequency band.

[0074] The component health includes the spring health and the valve core seal health.

[0075] Specifically, the deviation of dynamic features closely related to specific fault mechanisms is transformed into a quantitative description of component health status through mathematical models. When the coupled response analysis module identifies the current response mode as spring fatigue mode, the system initiates a health calculation process for the pressure regulating spring. This process focuses on the deviation of two core dynamic features. First, it quantifies the oscillation of the pressure recovery time, which can be achieved by calculating the variance or standard deviation of the outlet pressure time series in the excitation response data during the recovery phase; a larger variance value indicates more severe oscillation. Second, the system calculates the degree of anomaly in the vibration spectrum change within a preset low-frequency band, which can be obtained by calculating the ratio or difference between the energy integral in this frequency band and the baseline value of the health status. Finally, these two quantitative indicators are combined through a pre-established weighted fusion model, which calibrates the weights based on a large amount of experimental data, and finally outputs a normalized value, which is the spring health status. When the response mode is identified as valve core wear mode, the system calls the health calculation process for the valve core and valve seat sealing performance. This process also focuses on two core dynamic features. First, it accurately calculates the overshoot amplitude of the pressure recovery time, i.e., the difference between the maximum value reached by the outlet pressure curve during the recovery phase and the upper limit of the preset stable state range. A larger overshoot amplitude directly reflects the severity of the seal failure. Second, the system quantifies the degree of anomaly in the high-frequency band of the vibration spectrum change, especially the amplitude or energy of newly emerging spectral peaks. This requires differential comparison of the vibration spectrum during excitation with the healthy baseline spectrum, and identification and energy calculation of new spectral peaks. Similarly, the overshoot amplitude and the degree of high-frequency anomaly are calculated using a specific weighted fusion model to obtain a comprehensive quantitative value, which is the valve core seal health. Finally, the component health calculated in this step is a vector containing multiple sub-items, including spring health and valve core seal health. These specific health values, pointing to specific components, provide refined input for the fusion calculation of the next-level health index.

[0076] Optionally, the method further includes:

[0077] The health index is tracked over time to generate the rate of change of the health index;

[0078] When the health index is lower than a preset safety threshold or the rate of change of the health index is less than zero, an early warning signal is generated;

[0079] An early warning report is generated based on the warning signal and the corresponding health status of the component.

[0080] Specifically, firstly, the system performs time-series tracking on the periodically calculated health index. This means that after each complete monitoring and evaluation process, the newly generated health index is recorded and, together with historical data points, forms a time series reflecting the evolution of the equipment's health status over time. Based on this time series, the system calculates the rate of change of the health index. A simple calculation method is to use first-order backward difference to approximate the instantaneous rate of change: ,

[0081] in, It is the rate of change of the health index at time t. It is the health index at the current moment. It is the health index of the previous monitoring period. This is the time interval of the monitoring cycle. Next, the system will initiate the logic for generating early warning signals. This logic is based on two parallel judgment criteria. The first criterion is a state threshold judgment, where the system uses the latest health index... It is compared to a pre-set safety threshold. This safety threshold is a minimum acceptable health level set based on equipment importance, safety specifications, and maintenance experience. When the health status falls below this safety threshold, it indicates that the device's health has deteriorated to an unacceptable level, and the system immediately generates a warning signal. The second criterion is trend judgment; the system will determine the rate of change of the health index. Is it less than zero? A rate of change that is consistently less than zero indicates that the device's health is in an irreversible downward trend, foreshadowing future risks even if the current health index is still above the safe threshold. Therefore, when When the health index is less than zero, especially for several consecutive periods, the system will also generate a warning signal. The parallel use of these two criteria ensures dual coverage for both sudden severe failures and gradual performance degradation. Finally, once a warning signal is triggered, the system automatically generates a detailed warning report. This report is not just a simple alarm notification; it integrates all key information related to the warning. The report first clearly states the reason for the warning: whether the health index is below the threshold or the rate of change is continuously negative. More importantly, the report includes component health information corresponding to the current response mode; for example, it clearly indicates whether the spring health is too low or the valve core seal health is declining too rapidly. In this way, the warning report provides maintenance personnel with clear diagnostic conclusions and fault location, guiding them to conduct targeted inspections and repairs. Figure 3As shown, long-term tracking of the health index reveals a trend of degradation in the pressure reducer's health. For example, in the fifth month, a significant drop in the health index occurred due to the detection of a "spring fatigue mode." When the health index curve declines and crosses a preset safety threshold, such as 70, the system automatically generates an early warning signal, reminding management personnel to perform timely maintenance. This condition-based predictive maintenance enhances the intelligence level of equipment management and the system's safety.

[0082] Based on the same inventive concept, such as Figure 4 As shown, the present invention also provides a safety monitoring system for a gas pressure regulator, the system comprising:

[0083] The multi-source sensing module is used to acquire the outlet pressure, flow rate, valve body temperature and vibration signal of the gas pressure regulator to obtain multi-source monitoring data;

[0084] The operating condition excitation module is used to apply operating condition excitation to the gas pressure regulator and simultaneously collect the changes of the multi-source monitoring data during the excitation period to generate excitation response data.

[0085] The dynamic feature extraction module is used to extract pressure recovery time, vibration spectrum change and temperature change gradient from the excitation response data and combine them into a dynamic feature set;

[0086] The coupling analysis module is used to perform coupling response analysis on the dynamic feature set to identify response patterns that characterize the health status of internal mechanical components.

[0087] The health assessment module is used to quantitatively assess the health status of the gas pressure regulator based on the response pattern and generate a health index.

[0088] To verify the feasibility of this invention in practice, it was applied to a high-purity specialty gas supply system. Traditional periodic replacement and maintenance strategies are costly and cannot predict unforeseen failures. This factory aims to use the method of this invention to perform online health monitoring of critical gas pressure regulators, achieving a shift from reactive to predictive maintenance.

[0089] In this embodiment, high-precision pressure and flow sensors are installed on the outlet pipeline of the pressure reducer at the factory; thermocouple temperature sensors and broadband acceleration vibration sensors are installed on the valve body housing. These sensors are connected to a multi-source sensing module for synchronous data acquisition. Downstream of the pressure reducer, a miniature solenoid valve controlled by an operating condition excitation module is connected in series to actively apply flow disturbances. The system is set to automatically execute the monitoring process every 8 hours.

[0090] During the initial operation phase, specifically the first month, the pressure regulator was in a healthy state. After the system was subjected to operating condition excitation, the collected excitation response data showed that the outlet pressure quickly and smoothly recovered to the set value after the disturbance ended, with a pressure recovery time of 0.45 seconds, without overshoot or oscillation; the energy change of the vibration signal during excitation, i.e., the vibration spectrum change, was very small and mainly distributed in the random noise frequency band; the valve body temperature gradient was close to zero. The dynamic feature extraction module combined these features into a dynamic feature set, and the coupling analysis module matched it with a pattern library, identifying it as a "healthy mode." Based on this, the health assessment module generated a health index of 97.

[0091] In its fifth month of operation, the system detected significant changes in its dynamic characteristics. During a routine monitoring session on a certain day, the system recorded an extended pressure recovery time of 1.1 seconds, with noticeable low-frequency oscillations occurring during the outlet pressure recovery process. Simultaneously, the dynamic feature extraction module calculated a significant increase in the energy of the vibration spectrum changes in the low-frequency range (5-20Hz). The coupling analysis module, by analyzing the correlation between the extended pressure recovery time accompanied by oscillations and the increased energy of the vibration spectrum in the low-frequency range, identified the response mode as a "spring fatigue mode." The health assessment module first mapped this mode to an initial index range of [50-70], then calculated the spring health as 62% based on the amplitude of the pressure oscillations and the degree of deviation of the low-frequency vibration energy, ultimately generating a health index of 65.

[0092] Because the health index (65) was lower than the preset safety threshold (70) and the rate of change of the health index remained negative, the system immediately generated a warning signal. The warning report clearly stated: "Spring fatigue mode detected, current health index is 65, maintenance is recommended, with a focus on checking the pressure regulating spring." The maintenance personnel opened the cover for inspection according to the report and found that the pressure regulating spring did indeed show visible elastic potential energy decay.

[0093] To further verify the system's ability to identify different faults, a valve core wear fault was simulated on another experimental pressure reducer. System monitoring revealed an 18% pressure overshoot during the pressure recovery process of this pressure reducer, and simultaneously, a new spectral peak appeared in the high-frequency range (2-5kHz) of the vibration spectrum. The system identified this coupling feature as a "valve core wear mode" and calculated the valve core seal health to be 45%, with a final health index of 48, triggering an emergency alarm.

[0094] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0095] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A safety monitoring method for a gas pressure reducer, characterized by, The method comprises: acquiring the outlet pressure, flow rate, valve body temperature and vibration signal of the gas pressure reducer to obtain multi-source monitoring data; applying a working condition excitation to the gas pressure reducer and synchronously collecting changes in the multi-source monitoring data during the excitation to generate excitation response data; wherein the working condition excitation includes generating a flow disturbance by a micro electromagnetic valve arranged downstream of the gas pressure reducer; during the application of the working condition excitation, the changes in the multi-source monitoring data are collected at a preset sampling rate to obtain the excitation response data; extracting the pressure recovery time, vibration spectrum change and temperature change gradient from the excitation response data to combine into a dynamic feature set; performing coupled response analysis on the dynamic feature set to identify a response mode representing the health status of internal mechanical components; based on the response mode, quantitatively evaluating the health status of the gas pressure reducer to generate a health index.

2. The method of claim 1, wherein the method further comprises: The multi-source monitoring data comprises: collecting the outlet pressure by a pressure sensor, collecting the flow rate by a flow sensor, collecting the valve body temperature by a temperature sensor, and collecting the vibration signal by a vibration sensor; time-synchronously aligning the outlet pressure, flow rate, valve body temperature and vibration signal to combine into multi-source monitoring data.

3. The method of claim 1, wherein the method further comprises: The combination into a dynamic feature set comprises: calculating the time required for the outlet pressure to recover to a preset stable state range from the end of the excitation to obtain the pressure recovery time; calculating the energy change of the vibration signal during the excitation relative to before the excitation by spectrum analysis to obtain the vibration spectrum change; calculating the change rate of the valve body temperature during the excitation to obtain the temperature change gradient; combining the pressure recovery time, vibration spectrum change and temperature change gradient into a dynamic feature set.

4. The method of claim 3, wherein the method further comprises: The coupled response analysis on the dynamic feature set to identify a response mode representing the health status of internal mechanical components comprises: analyzing the correlation between the pressure recovery time and the vibration spectrum change to distinguish changes in system damping characteristics and abnormal mechanical impact; based on the correlation, matching the dynamic feature set with a preset mode library representing mechanical health status to identify the response mode.

5. The method of claim 4, wherein the method further comprises: The response mode comprises a health mode, a spring fatigue mode and a valve core wear mode; wherein the identification of the spring fatigue mode is based on the correlation that the pressure recovery time is prolonged and oscillates, and the vibration spectrum change is continuously abnormal in the low frequency band; the identification of the valve core wear mode is based on the correlation that a pressure overshoot appears in the pressure recovery time, and a new spectrum peak appears in the vibration spectrum change in the high frequency band.

6. The method of claim 5, wherein the method further comprises: The quantitative evaluation of the health status of the gas pressure reducer based on the response mode to generate a health index comprises: mapping the identified response mode to a preset health index value range to generate an initial index; calculating the health degree of the mechanical component according to the deviation degree of the dynamic feature in the response mode to obtain a component health degree; fusing the component health degree and the initial index to obtain a health index.

7. The method of claim 6, wherein the method further comprises: The health degree of the mechanical component is calculated according to the deviation degree of the dynamic characteristics in the response mode, and the component health degree is obtained. When the response mode is a spring fatigue mode, the spring health degree is calculated based on the oscillation degree of the pressure recovery time and the abnormal degree of the vibration frequency spectrum change in the low frequency band. When the response mode is a valve core wear mode, the valve core seal health degree is calculated based on the overshoot amplitude of the pressure recovery time and the abnormal degree of the vibration frequency spectrum change in the high frequency band. The component health degree includes the spring health degree and the valve core seal health degree.

8. The method of claim 6, wherein the method further comprises: The method further includes: tracking the health degree index in time series to generate a health degree index change rate; generating a warning signal when the health degree index is lower than a preset safety threshold or the health degree index change rate is less than zero; generating a warning report based on the warning signal and the corresponding component health degree.

9. A safety monitoring system for a gas pressure reducer, characterized by comprising: The system includes: A multi-source sensing module is configured to obtain the outlet pressure, flow rate, valve body temperature and vibration signal of the gas pressure reducer to obtain multi-source monitoring data. A working condition excitation module is configured to apply a working condition excitation to the gas pressure reducer and synchronously collect changes in the multi-source monitoring data during the excitation to generate excitation response data; wherein the working condition excitation includes generating a flow disturbance by a micro electromagnetic valve arranged downstream of the gas pressure reducer; during the application of the working condition excitation, the changes in the multi-source monitoring data are collected at a preset sampling rate to obtain the excitation response data. A dynamic characteristic extraction module is configured to extract the pressure recovery time, vibration frequency spectrum change and temperature change gradient from the excitation response data to combine them into a dynamic characteristic set. A coupling analysis module is configured to perform coupling response analysis on the dynamic characteristic set to identify a response mode representing the health state of the internal mechanical component. A health assessment module is configured to quantitatively assess the health state of the gas pressure reducer based on the response mode to generate a health degree index.

Images