A method for intelligent monitoring and abnormal early warning of the running state of a water quality detection device

Abstract

This invention belongs to the field of water quality testing equipment monitoring technology. It provides an intelligent monitoring and anomaly early warning method for the operating status of water quality testing equipment, including: continuously collecting historical action time data and corresponding historical ambient temperature data of pneumatic valves under historical clean conditions; establishing a health baseline model; and real-time collecting real-time action time data and corresponding real-time ambient temperature data of pneumatic valves under operating conditions; based on the temperature change rate of the real-time ambient temperature data. By transforming the inherent breathing effect of the equipment into a quantitative indicator for diagnosing gas source pollution, and combining it with adaptive period division, multi-dimensional feature extraction, trend analysis, and collaborative judgment, this invention achieves online monitoring and early warning of the gradual accumulation process of gas source pollution. It effectively solves the problem of normal temperature fluctuations masking abnormal pollutant characteristics, improving the accuracy and foresight of water quality testing equipment operating status monitoring.

Description

Technical Field

[0001] This invention belongs to the field of water quality testing equipment monitoring technology, specifically a method for intelligent monitoring and early warning of abnormalities in the operating status of water quality testing equipment. Background Technology

[0002] Water quality testing equipment is a key infrastructure in environmental protection and water resource management, and is widely used in automatic surface water monitoring stations and online pollution source monitoring systems. Such equipment usually contains a large number of pneumatic actuators, such as pneumatic valves that drive flow path switching. The reliability and accuracy of their operation directly determine the quality of water quality analysis data. The "operating status" of a valve not only refers to the realization of its on / off function, but also includes its dynamic response characteristics, such as the "action time" required for the valve to complete the action from receiving the command. Under ideal conditions, due to the thermal expansion and contraction characteristics of metal materials, the valve's action time will show regular small fluctuations with the periodic changes in ambient temperature. This phenomenon is figuratively called the "breathing effect" of the equipment, which is a potential indicator for measuring the mechanical health of the valve. However, in actual working conditions, the compressed air source that powers the equipment often contains trace amounts of water vapor and oil mist due to improper pretreatment. These contaminants will gradually deposit in the micron-level fit gap between the valve core and valve body of the pneumatic valve, forming viscous sludge. Currently, we usually only focus on whether the pressure of the air source meets the standard, or only carry out passive maintenance after the valve is completely stuck and the equipment reports an "overtime fault". There is a lack of online monitoring methods for the gradual accumulation process of contaminants. Furthermore, due to the objective existence of the diurnal temperature difference and the uneven length of day and night caused by seasonal changes, the valve's operating time itself is in dynamic fluctuation. This normal fluctuation caused by environmental factors can easily mask the abnormal characteristics caused by the aggravation of pollution. For example, maintenance personnel or traditional monitoring systems often simply attribute the overall slowdown in action time to temperature changes, failing to identify the amplified "breathing effect" caused by the sludge damping effect and the resulting distortion of dynamic response characteristics. This monitoring blind spot leads to the long-term lurking of air pollution problems, not only causing valve action delays and metering inaccuracies, resulting in a continuous decline in the repeatability and accuracy of water quality monitoring data, but also potentially triggering sudden equipment shutdowns without warning, seriously affecting the capture rate and effectiveness of water quality monitoring data, and posing significant hidden dangers to environmental management and decision-making. Therefore, the present invention provides a method for intelligent monitoring and early warning of abnormalities in the operating status of water quality testing equipment. Summary of the Invention

[0003] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0004] The technical solution adopted by this invention to solve its technical problem is: a method for intelligent monitoring and abnormal early warning of the operating status of water quality testing equipment, comprising: Continuously collect historical action time data and corresponding historical ambient temperature data of pneumatic valves under historical clean conditions, establish a health baseline model, and collect real-time action time data and corresponding real-time ambient temperature data of pneumatic valves under operating conditions. Based on the temperature change rate of real-time ambient temperature data, the inflection point of the temperature curve is automatically identified, and the time interval between adjacent temperature inflection points is used as the complete temperature fluctuation cycle to adaptively divide the analysis period. Within each temperature fluctuation cycle, the real-time action time data is decomposed to separate the periodic component caused by the periodic temperature change, the trend component caused by the accumulation of pollutants, and the residual component caused by random disturbance. The response characteristic curve of the real-time action time data relative to the real-time ambient temperature data is constructed, and the characteristic parameters of the response characteristic curve are extracted. Based on the state feature parameters extracted from the current temperature fluctuation cycle, the health deviation from the baseline features in the health baseline model is calculated, and the trend analysis of the health deviation is performed. Based on the results of health deviation and trend analysis, it is determined whether gas source pollution affects valve performance. If it does, an abnormal warning signal is issued.

[0005] The beneficial effects of this invention are as follows: This invention transforms the inherent breathing effect of the device into a quantitative indicator for diagnosing gas source pollution. Combined with adaptive period division, multi-dimensional feature extraction, trend analysis and collaborative judgment, it realizes online monitoring and early warning of the gradual accumulation process of gas source pollution. It effectively solves the problem that normal fluctuations caused by temperature mask the abnormal characteristics of pollutants, and improves the accuracy and foresight of water quality testing equipment operation status monitoring. This invention automatically identifies the inflection points of temperature curves based on real-time temperature change rate, and uses the time interval between adjacent inflection points as a complete temperature fluctuation cycle for data analysis. It can automatically adapt to changes in day and night length in different seasons and latitude regions, ensuring that each analysis cycle has physical integrity. This fundamentally eliminates the data comparability problem caused by differences in cycle length, and provides a time benchmark for subsequent feature extraction and trend analysis. Attached Figure Description

[0006] The invention will now be further described with reference to the accompanying drawings.

[0007] Figure 1 This is a flowchart of the steps of an intelligent monitoring and anomaly early warning method for the operating status of water quality testing equipment according to the present invention; Figure 2This is a block diagram of an intelligent monitoring and early warning system for the operating status of a water quality testing device according to the present invention. Detailed Implementation

[0008] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments. Example

[0009] One of the core inventive points of this invention is: addressing the shortcomings of current technology, a method for intelligent monitoring and anomaly early warning of water quality testing equipment operation status is proposed. By adaptively dividing the temperature fluctuation cycle based on the rate of change of ambient temperature, the interference of uneven day and night duration on monitoring data is eliminated. Furthermore, the valve action time is decoupled by feature extraction to extract trend components related to pollution accumulation, as well as multi-dimensional features such as temperature sensitivity coefficient and hysteresis loop area characterizing the dynamic response characteristics of the valve. This transforms the inherent "breathing effect" of the equipment into a quantitative indicator for diagnosing gas source pollution, optimizes the problem of normal fluctuations caused by temperature masking abnormal characteristics of pollutants, and can identify performance degradation caused by gas source pollution before the valve is completely stuck, thus improving the accuracy and foresight of water quality testing equipment operation status monitoring. Please see Figure 1 As shown in the embodiment of the present invention, a method for intelligent monitoring and abnormal early warning of the operating status of water quality testing equipment is applied to the pneumatic valve of water quality testing equipment, and includes the following steps: Step 1: Continuously collect historical action time data and corresponding historical ambient temperature data of pneumatic valves under historical clean conditions, establish a health baseline model, and collect real-time action time data and corresponding real-time ambient temperature data of pneumatic valves under operating conditions. The specific process of step one is as follows: First, it is necessary to define the time window of the historical clean status. Select the second to fourth week after the equipment is put into operation, or the third to tenth day of continuous operation after the pneumatic system has been thoroughly cleaned, the valve core has been replaced, and the air source precision filter has been replaced as the clean status data collection period. During this period, data from no less than 500 action cycles are acquired, with each valve action as the trigger event. The recorded content includes: the action time of each action, the corresponding timestamp, and the ambient temperature collected by the temperature sensor. The collected historical data were divided into intervals based on ambient temperature to establish a baseline database, including: Each degree Celsius is considered a temperature range (e.g., from -10℃ to 40℃, with each 1℃ representing a range). The mean and standard deviation of the action time within each ambient temperature range are statistically analyzed to form a temperature-action time baseline response curve, which serves as the action time-temperature response baseline. Extract the daily ambient temperature fluctuation range (i.e., the difference between the highest and lowest temperatures of the day) and the daily action time fluctuation range (i.e., the difference between the maximum and minimum values ​​of all action times of the day). Calculate the change in action time caused by a unit temperature change, i.e., the ratio of the temperature fluctuation range to the action time fluctuation range. Calculate the average value for all historical cleaning days to obtain the benchmark value for the change in action time caused by a unit temperature change, which serves as the baseline for the daily fluctuation characteristics. The action time data under historical clean conditions are arranged in chronological order to form a time series. The seasonal trend decomposition method (STL decomposition) is used to decompose it into three components: a part reflecting the regular fluctuation caused by the periodic change of diurnal temperature (periodic component), a part reflecting the long-term slow change trend (trend component), and a random noise part (residual component). The typical fluctuation amplitude of the periodic component after decomposition (e.g., the average difference between the daily maximum and minimum values ​​of the periodic component) and the noise level of the residual component (e.g., the standard deviation of the residual series) are recorded as the baseline of the time series decomposition. After the equipment is running normally, it is monitored in real time. Each time the valve is activated, the current activation time, the moment of activation, and the ambient temperature are automatically recorded to form a real-time data record of "activation moment - activation time - ambient temperature".

[0010] Step 2: Based on the temperature change rate of real-time ambient temperature data, automatically identify the inflection point of the temperature curve, and use the time interval between adjacent temperature inflection points as the complete temperature fluctuation cycle. Adaptively divide the analysis time period to eliminate the difference in cycle length caused by uneven day and night duration. The specific process of step two is as follows: continuously collect ambient temperature data to ensure that the process of day and night temperature change can be captured, and smooth the ambient temperature data to eliminate occasional sensor noise or instantaneous interference, so as to obtain a smoothed temperature time series. Based on the temperature time series, the rate of change of temperature over time is calculated point by point. Specifically, the temperature value of the previous moment is subtracted from the temperature value of the previous moment, and then divided by the time interval between the two moments to obtain the rate of change of temperature at that moment. Traverse the entire temperature sequence to identify the points where the temperature change trend reverses, i.e., the inflection points of the temperature curves. An inflection point must satisfy one of the following conditions: The inflection point from rising to falling: the rate of temperature change was positive (heating) at the previous moment, and negative (cooling) at the current moment, with the inflection point passing through zero. The inflection point from decreasing to increasing: the rate of temperature change was negative (decreasing) at the previous moment, and positive (increasing) at the current moment, passing through zero in between; It is worth noting that a temperature change rate threshold (e.g., ±0.2 degrees Celsius per hour) is set. Only when the temperature change rate changes from positive to negative or from negative to positive, and the absolute value of the change rate exceeds the temperature change rate threshold, is it considered a valid inflection point. All identified temperature inflection points are arranged in chronological order, and the time interval between two adjacent inflection points is a complete temperature fluctuation cycle. Each temperature fluctuation cycle includes a complete heating phase (from the lowest temperature point to the highest temperature point) and a cooling phase (from the highest temperature point to the lowest temperature point), or vice versa. Therefore, the length of the temperature fluctuation cycle is determined by the actual temperature change, rather than a fixed 24 hours, thus achieving adaptive adjustment to the unevenness of day and night length. For each identified temperature fluctuation cycle, record the start time and end time. The start time is the moment when the first temperature inflection point of the cycle occurs, and the end time is the moment when the last temperature inflection point of the cycle occurs (i.e., the start time of the next cycle). Each temperature fluctuation cycle is treated as an independent data analysis unit, outputting the start time, end time, cycle length (hours), temperature time series, and corresponding valve action time data. Step two enables dynamic adaptive segmentation of the analysis period based on the real-time temperature change rate, thereby automatically adapting to changes in day and night length in different seasons and latitudes, ensuring the physical integrity of each analysis cycle, and providing a unified and comparable time benchmark for subsequent "breathing effect" feature extraction.

[0011] Step 3: Within each temperature fluctuation cycle, decompose the real-time action time data to separate the periodic component caused by the periodic temperature change, the trend component caused by the accumulation of pollutants, and the residual component caused by random disturbances. Construct the response characteristic curve of the real-time action time data relative to the real-time ambient temperature data and extract the characteristic parameters of the response characteristic curve. The specific process of step three is as follows: for any temperature fluctuation cycle; Extract the corresponding timestamp sequence, valve action time sequence, and ambient temperature sequence; The acquired data sequence is preprocessed, and the action time series is decomposed using the STL decomposition method. STL decomposition is a robust time series decomposition method that can decompose a time series into three components with clear physical meaning: Periodic component: Reflects the regular fluctuations caused by the periodic changes in temperature between day and night. The waveform of this component should be highly correlated with the temperature change curve within the period, and its fluctuation amplitude characterizes the normal response strength of the valve to temperature. Trend component: Reflects the long-term trend of change after removing periodic fluctuations within the cycle. For a single-day cycle, the trend component usually appears as an approximately horizontal straight line. If there is a significant monotonic increase, it indicates that the valve action time continues to increase within the cycle, which may indicate an increase in basic resistance due to the accumulation of contaminants. Residual components: Reflect the random fluctuations that cannot be explained by periods and trends. The dispersion of the residuals (such as standard deviation) reflects the stability of valve operation. Among them, the terminal value of the trend component will serve as one of the bases for judging the pollution accumulation trend; Using the ambient temperature of the temperature fluctuation cycle as the horizontal axis and the corresponding action time as the vertical axis, all data points are plotted on a two-dimensional coordinate system. The data points will naturally form two branches: the heating segment branch (data points collected during the process of temperature rising from low to high) and the cooling segment branch (data points collected during the process of temperature falling from high to low). Curve fitting was performed on the data points of the heating and cooling sections respectively, resulting in two smooth curves; Among them, the fitting method can be polynomial fitting (such as quadratic polynomial) or local weighted regression scatter smoothing method. The two curves together constitute the action time-temperature response characteristic curve within the temperature fluctuation cycle. Extract the characteristic parameters of the response curve from the constructed action time-temperature response characteristic curve; Among them, the characteristic parameters include at least one of the following: the change in action time caused by a unit temperature change, the hysteresis loop area of ​​the heating and cooling curves, and the phase difference between the extreme values ​​of action time and temperature. The change in action time caused by a unit temperature change is used as the temperature sensitivity coefficient for the current cycle. The specific calculation method is as follows: The difference between the fitted action time value corresponding to the highest temperature point and the fitted action time value corresponding to the lowest temperature point within the cycle is divided by the temperature range within the cycle. Among them, the temperature sensitivity coefficient reflects the sensitivity of the valve action time to temperature changes under the current state. When the gas source pollution intensifies, the viscosity-temperature characteristics of the sludge will significantly increase this coefficient. The hysteresis loop area of ​​the heating and cooling curves. The hysteresis loop occurs because the viscoelasticity of contaminants (sludge) causes inconsistencies in the valve's frictional characteristics during the heating and cooling processes. The specific calculation method is as follows: Within the temperature range, the vertical distance between the heating section curve and the cooling section curve is integrated; The size of the hysteresis loop area directly reflects the severity of pollutant accumulation. The larger the area, the more significant the difference between the heating and cooling processes, and the more severe the pollution. The phase difference between the extreme values ​​of action time and temperature is analyzed to determine the phase relationship between the action time series and temperature series within the cycle. Under clean conditions, the lowest point of action time is usually basically synchronized with the highest point of temperature (or with a slight delay). The specific calculation method is as follows: The time delay between the action time series and the temperature series is analyzed using the cross-correlation function. The peak position of the cross-correlation function corresponds to the optimal alignment offset of the two series. Divide the optimal alignment offset by the period length to convert it into phase difference (in degrees or radians). When gas source pollution intensifies, due to the damping effect of oil sludge, the response time to temperature changes will show a significant lag, that is, the lowest point of the response time occurs significantly later than the highest point of the temperature, and the phase difference increases. The original feature parameters extracted for each temperature fluctuation cycle, including the temperature sensitivity coefficient, hysteresis loop area, phase difference, and the terminal value of the trend component and the residual noise level obtained from STL decomposition, are stored together as the state feature vector for that temperature fluctuation cycle.

[0012] Step 4: Based on the state feature parameters extracted from the current temperature fluctuation cycle, calculate the health deviation relative to the baseline feature and perform trend analysis of the health deviation; The specific process of step four is as follows: compare the end value of the trend component obtained by STL decomposition with the benchmark level of the trend component in the same period in the health baseline model, and obtain the pollution accumulation index, which characterizes the long-term accumulation of pollutants in the valve fitting clearance, by calculating the relative deviation between the two, reflecting the change in basic resistance caused by pollutant deposition. The temperature sensitivity coefficient is compared with the baseline value of the change in action time caused by a unit temperature change in the health baseline model. The relative change of the current temperature sensitivity coefficient relative to the baseline value is calculated as the increase of the temperature sensitivity coefficient, which reflects the degree of change in the valve's thermal response characteristics caused by the intervention of contaminants such as sludge. The hysteresis loop area extracted from the current temperature fluctuation cycle is compared with the baseline fluctuation amplitude of the temperature fluctuation cycle component in the healthy baseline model. The multiple relationship between the current hysteresis loop area and the baseline fluctuation amplitude is calculated. As the multiple of the hysteresis loop area, it reflects the degree of expansion of the difference in the dynamic response characteristics of the valve during the heating and cooling processes. The phase difference between the extreme value of the action time extracted from the current temperature fluctuation cycle and the extreme value of temperature is compared with the reference phase relationship of the synchronization between the two in the health baseline model. The absolute offset of the current phase difference relative to the reference phase difference is calculated as the phase difference offset, which reflects the degree of response hysteresis caused by the damping effect of pollutants. The current temperature fluctuation cycle yields health deviation indicators (i.e., pollution accumulation index, temperature sensitivity coefficient increase, hysteresis area multiple, and phase difference offset), which, together with the valve's historical deviation indicators for multiple consecutive temperature fluctuation cycles, constitute a multidimensional health deviation time series. It should be noted that the multidimensional health deviation time series records the evolution trajectory of pollution accumulation index, temperature sensitivity coefficient increase, hysteresis loop area multiple, and phase difference offset over time, providing a data basis for trend analysis. The number of historical periods selected should be sufficient to reflect the evolution pattern of equipment performance. Trend analysis based on multidimensional health deviation time series includes: The health deviation data of the most recent N consecutive temperature fluctuation cycles are selected as the analysis window. The value of N is determined according to the equipment operating frequency, usually 5 to 10 cycles. The window length should be sufficient to reflect the evolution trend of the indicator, while avoiding interference from historical data that is too far back in time. The Mann-Kendall nonparametric trend test method was used to conduct a statistical significance test on the changing trends of various health deviation indicators within the analysis window; For any health deviation index, the sequence value over N temperature fluctuation cycles; Calculate the standardized test statistic Z-value and the corresponding significance probability p-value, and the judgment rules are as follows: If the Z value is greater than 0 and the p value is less than 0.05, the indicator is considered to have a significant upward trend. If the Z value is greater than 0 and the p value is greater than or equal to 0.05, the indicator is considered to have a non-significant upward trend. If the Z value is less than or equal to 0, the indicator is considered to have no upward trend. For indicators that are determined to have a significant upward trend, further confirmation of the sustainability of the upward trend is needed: If the values ​​for the most recent three consecutive temperature fluctuation cycles are all higher than the values ​​for the previous temperature fluctuation cycle, it is judged as a continuous rise. If, within the analysis window, the value of no more than two temperature fluctuation cycles has fallen compared to the previous cycle, but the magnitude of the fall has not changed the overall upward trend, and the value of the cycle after the fall is still higher than the value of the initial cycle of the window, it is judged as a fluctuating increase. The trend judgment results of all health deviation indicators are summarized to form the trend feature vector of the current temperature fluctuation cycle, including: whether there is a significant upward trend; whether the upward trend is sustainable; and the percentage of items showing a significant upward trend. For example, suppose the pneumatic valve of a water quality testing device, after processing in steps one to three, has four health deviation indicators extracted within the most recent six consecutive temperature fluctuation cycles (cycle numbers T1 to T6), as shown in the table below: cycle Pollution accumulation index Temperature sensitivity coefficient increase Hysteresis loop area multiple Phase difference offset T1 1.02 1.01 times 1.03 times 0.5° T2 1.05 1.04 times 1.08 times 0.8° T3 1.12 1.09 times 1.15 times 1.3° T4 1.18 1.16 times 1.24 times 1.9° T5 1.25 1.24 times 1.35 times 2.6° T6 1.33 1.33 times 1.48 times 3.4° Using the most recent 5 periods (T2 to T6) as the analysis window, the Mann-Kendall trend test was performed on each of the four indicators: Pollution accumulation index: The sequence value continued to rise, and the test result Z=+2.81, p=0.005, which reached statistical significance and was judged to have a significant upward trend; moreover, each period from T2 to T6 was higher than the previous period, and it was judged to be a continuous monotonic increase. Temperature sensitivity coefficient increase: The sequence value continued to rise, and the test result Z=+2.81, p=0.005, which reached statistical significance and was judged to have a significant upward trend; and each period from T2 to T6 was higher than the previous period, which was judged to be a continuous monotonic increase. Hysteresis loop area multiple: The sequence value continued to rise, and the test result Z=+2.81, p=0.005, which reached statistical significance and was judged to have a significant upward trend; and each period from T2 to T6 was higher than the previous period, which was judged to be a continuous monotonically increasing trend. Phase difference offset: The sequence value continues to rise, and the test result Z=+2.81, p=0.005, which reaches statistical significance and is judged to have a significant upward trend; and each period from T2 to T6 is higher than the previous period, which is judged to be a continuous monotonically increasing trend. The above judgment results are summarized to form the trend feature vector of the current period T6: ; in, These represent whether there is a significant upward trend in the pollution accumulation index, the increase in the temperature sensitivity coefficient, the multiple of the hysteresis loop area, and the phase difference offset (1 indicates yes, 0 indicates no). "No" indicates whether the above four indicators have a continuous monotonically increasing characteristic (1 indicates yes, 0 indicates no); R represents the proportion of indicators with significant increases (a value between 0 and 1).

[0013] Step 5: Based on health deviation and trend analysis, determine whether gas source pollution affects valve performance. If it does, issue an abnormal warning signal. Step five involves obtaining the health deviation index and trend feature vector of the current temperature fluctuation cycle. Based on the mechanism by which gas source contamination affects valve performance, the following anomaly judgment rules are established: When the following conditions are met simultaneously, it is determined that gas source pollution has had a substantial impact on valve performance, which is an abnormal state, and an abnormal warning signal is issued. Condition 1: The current cycle's pollution accumulation index is 1.0, indicating that pollutant accumulation has caused the basic resistance to exceed the normal level under clean conditions; Condition 2: In the trend feature vector, at least three indicators show a significant increase marked as 1, and the corresponding continuous increase is marked as 1, indicating that multiple health deviation indicators show a continuous upward trend rather than random fluctuations. Condition 3: The proportion of significantly rising indicators is greater than or equal to 0.75, indicating that most health deviation indicators deteriorate synchronously and have a high degree of synergy; If the above three conditions are not met simultaneously, the valve is determined to be in a normal state. It should be noted that the values ​​in the above judgment rules (such as 1.0, 0.75, etc.) are exemplary settings. 1.0 corresponds to the baseline level of the health baseline and is a natural dividing line for distinguishing whether it deviates from the clean state. The requirement of three indicators (i.e. more than half) is based on the majority voting principle and is used to exclude accidental factors. The proportion of 0.75 corresponds to the synergistic characteristic of at least three of the four indicators deteriorating simultaneously. In practical applications, those skilled in the art can make adaptive adjustments to the above values ​​according to the specific equipment model, operating environment and historical statistical data to establish a judgment standard that conforms to the actual working conditions. After an abnormal warning is triggered, continuously monitor the health deviation indicators and trend characteristics of subsequent temperature fluctuation cycles: If the anomaly detection rules are consistently met in subsequent cycles, the alert status will be maintained. If the anomaly determination rules are no longer met in subsequent cycles and the system remains normal for three consecutive cycles, the warning will be automatically lifted and the event loop information will be recorded.

[0014] This embodiment explores the inherent breathing effect of pneumatic valves in a clean state, namely the physical phenomenon that the action time fluctuates periodically with the ambient temperature. By transforming this phenomenon, which was originally considered a normal fluctuation, into a quantitative indicator for diagnosing gas source pollution, an intrinsic correlation between pollutant accumulation and changes in the dynamic response characteristics of valves is established, providing a basis for the early identification of gas source pollution. This embodiment automatically identifies the inflection point of the temperature curve based on the real-time temperature change rate. It uses the time interval between adjacent inflection points as the complete temperature fluctuation cycle for data analysis. It can automatically adapt to the changes in day and night length in different seasons and latitude regions, ensuring that each analysis cycle has physical integrity. This fundamentally eliminates the data comparability problem caused by the difference in cycle length, and provides a time benchmark for subsequent feature extraction and trend analysis. This embodiment extracts multiple feature parameters from valve action time data, including trend components reflecting changes in basic resistance, temperature sensitivity coefficient characterizing thermal response characteristics, hysteresis loop area quantifying dynamic response differences, and phase difference characterizing response lag. It characterizes the impact of gas source pollution on valve performance from different physical perspectives, forming a multi-dimensional health assessment system that improves the accuracy and reliability of anomaly identification. This embodiment, by establishing a health baseline model, calculating health deviation indices, and conducting trend significance tests, can capture subtle changes in the gradual accumulation of contaminants in valve fitting clearances. When the contamination accumulation index exceeds the cleanliness benchmark level and multiple health deviation indices show a continuous and significant upward trend, performance degradation caused by gas source contamination can be identified in advance before the valve becomes completely stuck. This provides a sufficient time window for preventive maintenance of equipment and effectively reduces data interruptions caused by sudden equipment shutdowns. Example

[0015] Based on the same inventive concept as the intelligent monitoring and anomaly early warning method for the operating status of water quality testing equipment in the foregoing embodiments, such as Figure 2 As shown, this application provides an intelligent monitoring and anomaly early warning system for the operating status of water quality testing equipment, wherein the system specifically includes: Data acquisition module: continuously collects historical action time data and corresponding historical ambient temperature data of pneumatic valves under historical clean conditions, establishes a health baseline model, and collects real-time action time data and corresponding real-time ambient temperature data of pneumatic valves under operating conditions. The adaptive period division module automatically identifies the inflection point of the temperature curve based on the temperature change rate of real-time ambient temperature data. It uses the time interval between adjacent temperature inflection points as the complete temperature fluctuation period and adaptively divides the analysis period to eliminate the difference in period length caused by uneven day and night duration. Feature extraction module: Within each temperature fluctuation cycle, the real-time action time data is decomposed to separate the periodic component caused by the periodic temperature change, the trend component caused by the accumulation of pollutants, and the residual component caused by random interference. The response characteristic curve of the real-time action time data relative to the real-time ambient temperature data is constructed, and the feature parameters of the response characteristic curve are extracted. Health status assessment module: Based on the status feature parameters extracted from the current temperature fluctuation cycle, calculate the health deviation relative to the benchmark feature and perform trend analysis of the health deviation; Anomaly warning module: Based on health deviation and trend analysis, it determines whether gas source pollution affects valve performance. If it does, it issues an anomaly warning signal.

[0016] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent monitoring and early warning of abnormalities in the operating status of water quality testing equipment, characterized in that: include: Continuously collect historical action time data and corresponding historical ambient temperature data of pneumatic valves under historical clean conditions, establish a health baseline model, and collect real-time action time data and corresponding real-time ambient temperature data of pneumatic valves under operating conditions. Based on the temperature change rate of real-time ambient temperature data, the inflection point of the temperature curve is automatically identified, and the time interval between adjacent temperature inflection points is used as the complete temperature fluctuation cycle to adaptively divide the analysis period. Within each temperature fluctuation cycle, the real-time action time data is decomposed to separate the periodic component caused by the periodic temperature change, the trend component caused by the accumulation of pollutants, and the residual component caused by random disturbance. The response characteristic curve of the real-time action time data relative to the real-time ambient temperature data is constructed, and the characteristic parameters of the response characteristic curve are extracted. Based on the state feature parameters extracted from the current temperature fluctuation cycle, the health deviation from the baseline features in the health baseline model is calculated, and the trend analysis of the health deviation is performed. Based on the results of health deviation and trend analysis, it is determined whether gas source pollution affects valve performance. If it does, an abnormal warning signal is issued.

2. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 1, characterized in that: The specific process for establishing a health baseline model is as follows: Each degree Celsius is considered a temperature range. The mean and standard deviation of historical action time data within each ambient temperature range are statistically analyzed to form an action time-temperature response baseline. Extract the daily ambient temperature fluctuation range and the daily action time fluctuation range, calculate the change in action time caused by a unit temperature change, and calculate the average value of all historical cleaning days as the daily fluctuation characteristic baseline. The action time data under historical clean conditions are arranged in chronological order to form a time series. The seasonal trend decomposition method is used to decompose the data into periodic components, trend components, and residual components. The typical fluctuation amplitude of the periodic components and the noise level of the residual components are recorded as the baseline for time series decomposition.

3. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 1, characterized in that: The specific process of automatically identifying the inflection point of the temperature curve is as follows: The ambient temperature data is smoothed to obtain a smoothed temperature time series. The rate of change of temperature over time is calculated point by point. The position where the temperature change trend reverses is identified as the inflection point of the temperature curve. The inflection point includes the inflection point from heating to cooling and the inflection point from cooling to heating. All identified temperature inflection points are arranged in chronological order, and the time interval between two adjacent inflection points is considered as a complete temperature fluctuation cycle.

4. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 1, characterized in that: The specific process of decomposing real-time action time data is as follows: For any temperature fluctuation cycle, extract the corresponding timestamp sequence, valve action time sequence, and ambient temperature sequence; The STL decomposition method can be used to decompose the action time series into three components: periodic component, trend component, and residual component.

5. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 1, characterized in that: Extract the characteristic parameters of the response characteristic curve, including: The change in action time caused by a unit temperature change, the hysteresis loop area of ​​the heating and cooling curves, and at least one of the following: the phase difference between the extreme values ​​of action time and temperature.

6. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 5, characterized in that: The difference between the fitted action time value corresponding to the highest temperature point and the fitted action time value corresponding to the lowest temperature point within the cycle is divided by the temperature range within the cycle, i.e. the change in action time caused by a unit temperature change, and is used as the temperature sensitivity coefficient for the current cycle. Within the temperature range, the vertical distance between the heating segment curve and the cooling segment curve is integrated to calculate the hysteresis loop area of ​​the heating segment and the cooling segment curve. The time delay between the action time series and the temperature series is analyzed using the cross-correlation function. The peak position of the cross-correlation function corresponds to the optimal alignment offset of the two series. The optimal alignment offset is divided by the period length and converted into a phase difference, which is used as the phase difference between the extreme values ​​of the action time and the extreme values ​​of the temperature.

7. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 1, characterized in that: Based on the state characteristic parameters extracted from the current temperature fluctuation cycle, the health deviation relative to the baseline characteristics is calculated, including: pollution accumulation index, temperature sensitivity coefficient increase, hysteresis loop area multiple, and phase difference offset. The relative deviation is calculated by comparing the terminal value of the trend component with the benchmark level of the trend component in the same period of the healthy baseline model, and the pollution accumulation index is obtained. The temperature sensitivity coefficient is compared with the baseline value of the change in action time caused by a unit temperature change in the health baseline model, and the relative change amplitude is calculated to obtain the increase of the temperature sensitivity coefficient. The hysteresis loop area is compared with the baseline fluctuation amplitude of the periodic component in the healthy baseline model, and the multiple relationship is calculated to obtain the hysteresis loop area multiple. The phase difference is compared with the baseline phase relationship of synchronicity between the two in the healthy baseline model, and the absolute offset is calculated to obtain the phase difference offset.

8. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 1, characterized in that: The specific process for conducting health deviation trend analysis is as follows: The health deviation data of the most recent N consecutive temperature fluctuation cycles are selected as the analysis window; The Mann-Kendall nonparametric trend test method was used to conduct a statistical significance test on the changing trends of various health deviation indicators within the analysis window; for indicators that were determined to have a significant upward trend, the sustainability of the upward trend was further confirmed. The trend judgment results of all health deviation indicators are summarized to form the trend feature vector of the current temperature fluctuation cycle.

9. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 8, characterized in that: For indicators that are determined to have a significant upward trend, the process of further confirming the sustainability of the upward trend is as follows: Calculate the standardized test statistic Z-value and the corresponding significance probability p-value; If the Z-value > 0 and the p-value < 0.05, a significant upward trend is identified, further confirming the persistence of the upward trend. If the values ​​for the most recent three consecutive temperature fluctuation cycles are all higher than the values ​​for the previous temperature fluctuation cycle, it is judged as a continuous rise. If, within the analysis window, the value of no more than two temperature fluctuation cycles has fallen compared to the previous cycle, but the magnitude of the fall has not changed the overall upward trend, and the value of the cycle after the fall is still higher than the value of the initial cycle of the window, it is judged as a fluctuating increase.

10. The method for intelligent monitoring and early warning of abnormal operation status of water quality testing equipment according to claim 1, characterized in that: The specific process for issuing an abnormal warning signal is as follows: Obtain the health deviation index and trend feature vector of the current temperature fluctuation cycle; When the following conditions are met simultaneously, it is determined that gas source contamination has had a substantial impact on valve performance, constituting an abnormal state, and an abnormal warning signal is issued: Condition 1: The pollution accumulation index for the current cycle is greater than 1.0; Condition 2: In the trend feature vector, at least three indicators show a significant upward trend, and the corresponding increase is continuous; Condition 3: The proportion of significantly rising indicators is greater than or equal to 0.75.

Images