CN-121615521-B - Service life prediction method, system and storage medium of fire valve

CN121615521BCN 121615521 BCN121615521 BCN 121615521BCN-121615521-B

Abstract

The application relates to the field of fire safety operation and maintenance management, and discloses a service life prediction method, a service life prediction system and a storage medium of a fire valve. The method comprises the steps of collecting operation data of the fire valve, extracting degradation characteristics, analyzing interaction of the fire valve and environmental factors to construct a torque change rate model, determining nonlinear characteristic parameters based on fusion of the torque change indexes and the environmental influences, estimating linkage influences to obtain influence coefficients when the parameters exceed a threshold value, adjusting a seal coarse simulation grid and determining a dynamic evolution path, integrating connection strength of the operation frequency fluctuation update path to extract attenuation trend, determining degradation threshold value points, generating compensation signals to optimize the torque change rate model to obtain nonlinear rate correction values, quantifying cumulative effects based on the correction values, and dividing service life intervals of the fire valve. The application solves the problem that the performance degradation of the fire valve is difficult to accurately evaluate under the complex working condition, and improves the accuracy of life prediction and the scientificity of maintenance decision.

Inventors

SHEN LIANG
LI NAN
ZHANG XUYANG
LI XINGQI

Assignees

浙江中实安全科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260202

Claims (9)

1. The service life prediction method of the fire valve is characterized by comprising the following steps of: S101, collecting operation data of a fire valve, extracting degradation characteristics from the operation data, analyzing interaction between the degradation characteristics and environmental factors of a fire scene, constructing a torque change rate model and obtaining a torque change index; step S102, fusing the environmental influence of the fire scene according to the torque change index, and determining nonlinear characteristic parameters; Step S103, if the nonlinear characteristic parameter exceeds a preset characteristic threshold, evaluating linkage influence to obtain an influence coefficient; step S104, adjusting a seal rough simulation grid through the influence coefficient to determine a dynamic evolution path of the sealing surface of the fire valve, wherein the seal rough simulation grid is a digital discrete model for quantifying the rough characteristic of the sealing surface of the fire valve; step 105, integrating the operation frequency fluctuation data of the fire valve according to the dynamic evolution path, updating the connection strength between nodes in the dynamic evolution path, extracting the attenuation trend of the sealing performance, and determining the degradation threshold point; step S106, if the degradation threshold point is close to a preset alarm line, generating a compensation signal, and optimizing a torque change rate model through the compensation signal to obtain a nonlinear rate correction value; Step S107, based on the nonlinear rate correction value, adjusting related parameters of the linkage effect, quantifying the accumulated effect of seal roughness increase, deducing a final sealing performance index, and dividing a life remaining section of the fire valve according to the final sealing performance index; wherein, the step S104 includes: Multiplying the influence coefficient with the initial roughness value of each unit in the sealed rough simulation grid to realize the dynamic adjustment of the sealed rough simulation grid; Acquiring roughness change data of each unit in the adjusted sealed rough simulation grid in a set time step, and generating a roughness increment distribution map; Based on the mapping relation between the roughness increment distribution diagram and the torque change index, analyzing interaction between a multidimensional degradation factor and the roughness change of the sealing surface in a fire scene to obtain an evolution trend curve of the sealing surface of the fire valve; and identifying key nodes and change inflection points in the evolution process according to the nonlinear law of the evolution trend curve quantization sealing surface roughness changing along with time, and determining a dynamic evolution path of the fire valve sealing surface.
2. The method according to claim 1, wherein the step S101 includes: acquiring operation data of the fire valve under a fire scene through a sensor, wherein the operation data comprises operation frequency data and a torque signal sequence; performing wavelet transformation decomposition on the torque signal sequence to extract multidimensional degradation feature vectors; Quantifying interaction of the multidimensional degradation feature vector and fire scene environmental factors by adopting a time sequence analysis method, constructing a joint time sequence and establishing a torque change rate model; and calculating a differential slope of the joint time sequence based on the torque change rate model, and determining the differential slope as a torque change index.
3. The method according to claim 2, wherein the step S102 includes: calculating the influence weight of the fire scene environmental factors according to the torque change index; The torque change indexes are subjected to weighted fusion through the influence weights to form nonlinear rate data points and construct a nonlinear rate distribution curve; quantifying the corresponding relation between the nonlinear rate distribution curve and the mechanical wear degree of the fire valve by adopting a curve fitting technology; And extracting curvature, slope peak value and inflection point position from the corresponding relation to form a nonlinear characteristic parameter set.
4. The method according to claim 1, wherein the step S103 includes: if any parameter value in the nonlinear characteristic parameter set exceeds a corresponding preset characteristic threshold value, judging that the parameter value exceeds the corresponding preset characteristic threshold value; when the variance contribution rate exceeds the average value, evaluating linkage influence and calculating the variance contribution rate of each parameter in the nonlinear characteristic parameter set, and screening out the parameter with the variance contribution rate higher than the average value as a dominant component; Constructing a random forest model based on fire valve historical operation data, inputting the dominant component into the random forest model, and predicting an acceleration effect value of the dominant component on seal roughness growth; And according to the acceleration effect value and the influence weight of the fire scene environmental factors, obtaining an influence coefficient through weighted summation and logarithmic conversion calculation.
5. The method according to claim 1, wherein the step S105 includes: The dynamic evolution path is correspondingly used as a node sequence in a sealing surface roughness state, operation frequency fluctuation data of the fire valve are collected, and the operation frequency fluctuation data are overlapped to the node sequence to form integrated path data; Calculating roughness difference values between adjacent nodes based on the integrated path data, and taking the reciprocal of the roughness difference values as an initial value of the connection strength between the nodes in the dynamic evolution path; updating the connection strength between nodes in the dynamic evolution path by adopting an iterative optimization technology based on the integrated path data; Extracting continuous data sequences from the updated connection strength, and fitting the continuous data sequences by adopting a least square method to obtain a sealing performance attenuation trend; And comparing the sealing performance attenuation trend with historical degradation data of the fire valve, identifying an inflection point of the trend line from stable to abrupt drop, and determining the inflection point as a degradation threshold point.
6. The method according to claim 1, wherein the step S106 includes: calculating a numerical value difference value between the degradation threshold value point and a preset alarm line, and judging that the degradation threshold value point is close to the preset alarm line if the numerical value difference value is smaller than a preset difference value threshold value; When the time sequence segments are judged to be close, extracting time sequence segments from the attenuation trend of the sealing performance, performing frequency domain conversion on the time sequence segments, identifying periodic fluctuation components and generating a compensation signal sequence; Decoding the compensation signal sequence to obtain a target torque change rate, and calculating the absolute difference between the target torque change rate and the output value of the torque change rate model to obtain a calibration residual error; Adopting a particle swarm optimization algorithm to iteratively optimize parameters of the torque change rate model until the calibration residual error is lower than a preset residual error threshold value, and obtaining an optimized torque change rate model; and according to the optimized torque change rate model, combining the torque change indexes to obtain a nonlinear rate correction value.
7. The method according to claim 1, wherein the step S107 includes: Extracting an adjustment factor from the nonlinear rate correction value, and multiplying the adjustment factor by a related parameter of the linkage influence to obtain an adjusted linkage influence parameter; acquiring roughness accumulation and growth data of each unit of the adjusted seal rough simulation grid, and quantifying the accumulation effect of seal rough growth by combining the adjusted linkage influence parameters; deriving a final sealing performance index through a weighted fusion algorithm based on the cumulative effect and the nonlinear rate correction value; calculating the estimated time length of the final sealing performance index from the current value to a preset sealing performance failure threshold value based on the sealing performance attenuation trend; and dividing the life remaining section of the fire valve according to the estimated time length, wherein the dividing standard of the life remaining section is set based on the valve use requirement corresponding to the fire scene security level.
8. A service life prediction system for a fire valve, for implementing the service life prediction method for a fire valve according to any one of claims 1 to 7, wherein the service life prediction system for a fire valve comprises: The data acquisition module is used for acquiring the operation data of the fire valve, extracting degradation characteristics from the operation data, analyzing the interaction between the degradation characteristics and environmental factors of a fire scene, constructing a torque change rate model and obtaining a torque change index; The feature extraction module is used for fusing the environmental influence of the fire scene according to the torque change index and determining nonlinear feature parameters; the influence evaluation module is used for evaluating linkage influence when the nonlinear characteristic parameter exceeds a preset characteristic threshold value to obtain an influence coefficient; The evolution analysis module is used for adjusting a seal coarse simulation grid through the influence coefficient to determine a dynamic evolution path of the sealing surface of the fire valve, wherein the seal coarse simulation grid is a digital discrete model for quantifying the coarse characteristic of the sealing surface of the fire valve; the trend determining module is used for integrating the operation frequency fluctuation data of the fire valve according to the dynamic evolution path, updating the connection strength among nodes in the dynamic evolution path, extracting the attenuation trend of the sealing performance and determining a degradation threshold point; The model optimization module is used for generating a compensation signal when the degradation threshold value point approaches to a preset alarm line, and optimizing a torque change rate model through the compensation signal to obtain a nonlinear rate correction value; and the service life dividing module is used for adjusting related parameters of the linkage influence based on the nonlinear rate correction value, quantifying the accumulated effect of seal roughness increase, deducing a final sealing performance index, and dividing a service life remaining section of the fire valve according to the final sealing performance index.
9. A computer readable storage medium having stored thereon a computer program, which when run by a processor causes the processor to perform the method of life prediction of a fire valve according to any one of claims 1 to 7.

Description

Service life prediction method, system and storage medium of fire valve Technical Field The application relates to the field of fire safety operation and maintenance management, in particular to a service life prediction method, a service life prediction system and a storage medium of a fire valve. Background The fire valve is a key control part of fire water supply, gas fire extinguishing and other systems, and the reliability of the sealing performance of the fire valve is directly related to the effective transportation of fire extinguishing media and the safety of the system in fire. Under the severe working condition of long-term standby or sudden high-frequency opening and closing, the sealing surface of the fire-fighting valve is easy to degrade due to factors such as material aging, mechanical abrasion, environmental corrosion (such as humidity and chemical medium) and the like, and once the fire-fighting valve fails, the system is possibly paralyzed. Therefore, the method for accurately evaluating the state of the sealing performance of the fire-fighting valve and predicting the service life is an important technical requirement for realizing the operation and maintenance of the active fire-fighting equipment and guaranteeing the fire rescue capability. Along with the improvement of the intelligent level of the fire protection system, higher requirements are put on the health state monitoring and life prediction of key equipment such as fire protection valves. However, the technical scheme adopted currently still has various limitations, and is difficult to meet the accurate evaluation requirement under the complex fire scene, and the method is particularly characterized in that firstly, the data acquisition dimension is insufficient and multi-source information fusion is lacked, the existing method is mostly dependent on threshold value alarming or periodical manual inspection of a single sensor, and synchronous continuous monitoring and comprehensive analysis of multi-dimensional parameters such as torque dynamic change, opening and closing frequency fluctuation, environment temperature and humidity, medium corrosiveness and the like in the valve operation process are difficult to realize. Secondly, the interaction effect of the environment and the mechanical factors cannot be effectively modeled, most life assessment methods still adopt an empirical-based linear abrasion model or a static attenuation coefficient, and nonlinear dynamic coupling effect between the environment factors and the mechanical state of the valve is not fully considered. Thirdly, the prior art lacks an adaptive feedback and dynamic calibration mechanism, generally does not have the online optimization capability of a model based on real-time monitoring data, when performance indexes are close to a preset warning threshold value, a system cannot automatically generate compensation signals and reversely calibrate prediction model parameters, errors are accumulated along with time, the residual life interval is divided inaccurately, crossing from alarm to prediction is difficult to achieve, and timeliness and scientificity of preventive maintenance decisions are restricted. Fourth, the technology system is fragmented, a complete evaluation closed loop is not formed, most schemes still stay in a local function implementation stage, a systematic technology chain from data perception, feature extraction, interaction analysis and dynamic evolution modeling to life prediction and feedback optimization is not constructed, and the requirements of intelligent fire protection on the refinement and the intellectualization of the full life cycle health management of key equipment are difficult to meet. According to the application, by combining real-time operation data with multidimensional environment information acquisition, degradation characteristic interaction analysis, nonlinear dynamic evolution modeling and a feedback calibration mechanism based on performance threshold values, a closed-loop evaluation system which covers data perception, characteristic extraction, model optimization and life prediction is constructed, the problems that the sealing performance degradation of a fire valve is difficult to evaluate accurately and the life prediction reliability is insufficient under complex working conditions are solved, and the accuracy of state monitoring, the timeliness of early warning and the scientificity of maintenance decision are improved. Disclosure of Invention The application provides a service life prediction method, a service life prediction system and a storage medium of a fire valve, which solve the problems that the sealing performance of the fire valve is difficult to accurately evaluate under complex working conditions and the service life prediction reliability is insufficient, and improve the accuracy of state monitoring, the timeliness of early warning and the scientificity of maintenance decision. In a first