An evaluation method and system for the removal capacity of iron and manganese in underground seawater

Abstract

The application provides an underground seawater iron and manganese removal capacity evaluation method and system, relates to the technical field of underground seawater treatment, and comprises the following steps: collecting water quality index parameters of seawater before and after filtration; identifying a turbidity sudden increase event according to the relative change rate of turbidity and generating an event marker; identifying a pressure difference deterioration rate according to the change rate of the pressure difference between adjacent level filtration devices; inputting the iron removal amount and the manganese removal amount into a preset multi-factor removal capacity calculation model to obtain iron and manganese removal capacity index values; constructing a semi-parametric model to predict the residual service life proportion of filter material; introducing the pressure difference deterioration speed and the iron and manganese removal capacity index to quantify the performance attenuation trend of the system, thereby avoiding the interference of single instantaneous value fluctuation; and combining time sequence input features, i.e., the capacity index, the pressure difference speed and backwashing events, fusing data driving and physical laws, and being compatible with nonlinear attenuation characteristics.

Description

Technical Field

[0001] This invention relates to the field of underground seawater treatment technology, specifically to a method and system for evaluating the iron and manganese removal capacity of underground seawater. Background Technology

[0002] With the continuous development of water treatment technology, the requirements for water quality monitoring in fields such as seawater desalination and wastewater treatment are becoming increasingly stringent. Especially in the process of iron and manganese ion removal, the real-time monitoring of water quality indicators and treatment efficiency directly affect the performance and economy of the water treatment system. Traditional water quality monitoring methods often cannot provide sufficient real-time data, leading to a lag in the evaluation of filtration system performance, thereby affecting the overall efficiency of water treatment.

[0003] In the prior art, CN116797094A discloses a method for classifying water purification capacity by using evaluation data of various water quality indicators and based on a comprehensive evaluation index value. However, this method is limited to evaluating water quality and does not take into account the toxic effects of high concentrations of iron and manganese ions in water. High concentrations of iron and manganese ions can have adverse effects on the respiration, immunity, growth and development, and genotoxicity of aquatic organisms, thereby reducing their survival rate and growth rate. Therefore, it is particularly important to develop a system that can accurately monitor changes in water quality, automatically identify the performance degradation of filtration devices, and issue timely alarms.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for evaluating the iron and manganese removal capacity of underground seawater, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for assessing the iron and manganese removal capacity of underground seawater, comprising the following steps:

[0008] S1: Multiple consecutive monitoring cycles are preset. Water quality parameters of seawater before and after filtration are continuously collected in each monitoring cycle. The water quality parameters include the concentration of ferrous ions and the concentration of manganese ions. The pressure difference between adjacent filtration devices and the iron and manganese turbidity of the manganese sand filtration system are collected simultaneously and continuously in each monitoring cycle. The collected data are all aggregated according to the monitoring cycle.

[0009] S2: Based on the continuous data of iron and manganese turbidity within the monitoring period, calculate the relative change rate of turbidity in each monitoring period, identify turbidity mutations through the relative change rate of turbidity to generate backwash operation event markers, calculate the change rate of pressure difference between adjacent filter devices within the monitoring period based on the pressure difference data of adjacent filter devices within the monitoring period, and combine all the change rates to obtain the pressure difference deterioration rate.

[0010] S3: Calculate the difference between the influent and effluent concentrations of ferrous ions and manganese ions in the monitoring period to obtain the iron removal amount and manganese removal amount in the monitoring period. After obtaining the iron removal amount and manganese removal amount in each monitoring period, use them as input to the preset multi-factor iron and manganese removal capacity calculation model and output the iron and manganese removal capacity index of the corresponding monitoring period.

[0011] S4: The iron and manganese removal capacity index, backwashing operation event marker, and differential pressure deterioration rate of multiple consecutive monitoring cycles are used as time-series input features and input into the semi-parametric filter media life prediction model. The model outputs the predicted life percentage of the filter media in the corresponding monitoring cycle and uses the gradient descent method to dynamically correct the model parameters.

[0012] S5: Based on the predicted lifespan of the filter media and the iron and manganese removal capacity index, compare them in real time with the preset lifespan percentage threshold and capacity threshold respectively. When the iron and manganese removal capacity index is lower than the capacity threshold in a continuous monitoring period, or the predicted lifespan percentage of the filter media is lower than the lifespan threshold, an alarm signal is generated.

[0013] Furthermore, multiple consecutive monitoring cycles refer to dividing the current time into several consecutive time periods of equal length, with the current time as the endpoint. Each time period is defined as a monitoring cycle. When calculating the relative rate of change of turbidity in each monitoring cycle, the continuous data of iron and manganese turbidity are integrated over the monitoring cycle and normalized with the duration of the monitoring cycle to obtain the average turbidity value within the cycle.

[0014] The continuous data of iron and manganese turbidity within the monitoring period are compared with the average turbidity value within the monitoring period. Based on the relative change between the two, the rate of change of iron and manganese turbidity in the current monitoring period relative to the average turbidity in this monitoring period is calculated and calibrated as the relative turbidity change rate of the corresponding monitoring period.

[0015] The logic for identifying turbidity abrupt changes by the relative rate of change of turbidity is as follows: when the relative rate of change of turbidity reaches or exceeds a preset threshold, and the relative rate of change of turbidity reaches or exceeds the threshold determined by the product of the standard deviation of iron and manganese turbidity and the sensitivity coefficient within the corresponding monitoring period, when both of the above conditions are met, the event marker at that moment is set to "1", indicating that a turbidity abrupt increase event has occurred.

[0016] Furthermore, based on the pressure difference data of adjacent filtration units within the monitoring period, the rate of change of the pressure difference between adjacent filtration units within the monitoring period is calculated, and all rates of change are combined to obtain the pressure difference deterioration rate, specifically:

[0017] In the Within each monitoring period, the rate of change of pressure difference is obtained by differentiating the pressure difference data. When the rate of change of differential pressure is positive, record the rate of differential pressure deterioration:

[0018] ;

[0019] in, Indicates the first The rate of differential pressure deterioration within each monitoring cycle; Indicates the rate of change of pressure difference; The symbol representing the rate of change of pressure difference, when When the condition is met, the value is 1; otherwise, the value is 0. Indicates the first The start time within each monitoring cycle; Indicates the first The end time within each monitoring cycle.

[0020] Furthermore, the differences in influent and effluent concentrations of ferrous ions and manganese ions during the monitoring period were calculated to obtain the iron and manganese removal amounts for that period, specifically:

[0021] The differences in influent and effluent concentrations of ferrous and manganese ions within the monitoring period are calculated separately. The iron and manganese removal amounts at each moment in the monitoring period are integrated to obtain the total iron and manganese removal amounts for the entire monitoring period.

[0022] ;

[0023] ;

[0024] in, Indicates the first Iron removal amount within a monitoring cycle; Indicates the first Manganese removal amount within a monitoring cycle; Indicates the first Within each monitoring period Iron removal rate at any time; Indicates the first Within each monitoring period Manganese removal rate at any given time.

[0025] Furthermore, after obtaining the iron and manganese removal amounts for each monitoring period, these amounts are used as inputs to a pre-defined multi-factor iron and manganese removal capacity calculation model, which outputs the iron and manganese removal capacity index for the corresponding monitoring period. The specific steps are as follows:

[0026] Based on iron and manganese removal rates, a multi-factor removal capacity calculation model is constructed:

[0027] ;

[0028] in, Indicates the first The removal capacity index values ​​of iron and manganese within each monitoring cycle; Indicates the maximum iron removal capacity designed; Indicate the maximum manganese removal capacity designed; This represents the iron removal weighting coefficient; This represents the weighting coefficient for manganese removal, where , ,and .

[0029] Furthermore, the iron and manganese removal capacity indicators, backwashing operation event markers, and differential pressure deterioration rates from multiple consecutive monitoring periods are used as time-series input features and input into the semi-parametric filter media life prediction model. The specific output of the predicted lifespan percentage of the filter media for the corresponding monitoring period is as follows:

[0030] Constructing the filter media lifespan factor:

[0031] ;

[0032] in, Indicates the first Filter media lifespan factor within each monitoring cycle; This represents the theoretical optimal value for the number of backwashes per unit time. This represents the theoretically optimal value for the rate of pressure differential degradation. This represents the theoretical optimal value of the iron and manganese removal capacity index; This indicates the sensitivity coefficient to the impact of the number of backwashing cycles on the lifespan. This represents the sensitivity coefficient to the impact of differential pressure degradation rate on lifetime. This indicates the sensitivity coefficient to the impact of iron and manganese removal capacity on the product's lifespan. Indicates the first Number of backwashing operations within a monitoring cycle;

[0033] Using the filter media lifespan factor as a semi-parametric adjustment factor, a semi-parametric model is constructed to predict the proportion of filter media lifespan:

[0034] ;

[0035] in, Indicates the predicted first The percentage of remaining lifespan of the filter media in each monitoring cycle.

[0036] Furthermore, the gradient descent method is used to dynamically adjust the model parameters, specifically as follows:

[0037] Mean square error of the filter media lifespan influencing factors in the semi-parametric model:

[0038] ;

[0039] in, This represents the theoretical filter media lifespan factor value. The mean square error representing the loss of filter media influence factors;

[0040] Calculate separately for , , gradient:

[0041] ;

[0042] ;

[0043] ;

[0044] in, express for The gradient; express for The gradient; express for The gradient;

[0045] Parameter updates are performed using gradient descent.

[0046] ;

[0047] in, This represents the parameters before the update, where To affect the sensitivity coefficient index, , representing the sensitivity coefficient of the number of backwashing cycles to the lifespan. , representing the sensitivity coefficient to the impact of differential pressure degradation rate on lifetime. , representing the sensitivity coefficient of the iron and manganese removal capacity index to the lifespan; Indicates the updated parameters ; Indicates the learning rate; The loss function represents the... The gradient.

[0048] Further, an alarm signal is generated, and the specific steps are as follows:

[0049] Set the remaining lifespan threshold of the filter media as follows: Determine and calibrate the predicted remaining lifespan of the filter media for the most recent monitoring period. ,like Then, trace back... If the predicted remaining life of the filter media is less than [a certain value] for all monitoring cycles traced back, then [the following is a separate, unrelated sentence:] ...for each monitoring cycle, if the predicted remaining life of the filter media for all monitoring cycles traced back is less than [a certain value]... If the filter media life is insufficient, a warning signal will be issued; otherwise, no warning signal will be issued.

[0050] Set the capability threshold as Without issuing an early warning signal, determine and calibrate the iron and manganese removal capacity index of the monitoring period closest to the current time. ,like Then trace back... If the iron and manganese removal capacity index for all monitored periods is less than [a certain value], then [the following is a possible interpretation:] ... If so, an abnormal warning signal for iron and manganese removal capacity will be issued.

[0051] The present invention also provides a system for assessing the iron and manganese removal capacity of underground seawater, the system being used to perform the above-described assessment method, comprising:

[0052] The data acquisition module is used to preset multiple continuous monitoring cycles. In each monitoring cycle, it continuously collects water quality index parameters of seawater before and after filtration. The water quality index parameters include the concentration of ferrous ions and the concentration of manganese ions. Simultaneously, it continuously collects the pressure difference between adjacent levels of filtration devices and the iron and manganese turbidity of the manganese sand filtration system in each monitoring cycle, and collects the data according to the monitoring cycle.

[0053] The turbidity identification and differential pressure degradation calculation module is used to calculate the relative turbidity change rate of each monitoring period based on the continuous data of iron and manganese turbidity within the monitoring period. It generates backwash operation event markers by identifying turbidity abrupt changes through the relative turbidity change rate. Based on the pressure difference data of adjacent level filter devices within the monitoring period, it calculates the change rate of pressure difference between adjacent level filter devices within the monitoring period and integrates all change rates to obtain the differential pressure degradation rate.

[0054] The iron and manganese removal calculation module is used to calculate the difference between the influent and effluent concentrations of ferrous ions and manganese ions in the monitoring period, respectively, to obtain the iron and manganese removal amount in that monitoring period. After obtaining the iron and manganese removal amount in each monitoring period, it is used as input to be sent to the preset multi-factor iron and manganese removal capacity calculation model, and the iron and manganese removal capacity index of the corresponding monitoring period is output.

[0055] The filter media life prediction module takes the iron and manganese removal capacity index, backwash operation event markers, and differential pressure deterioration rate of multiple consecutive monitoring cycles as time-series input features, inputs them into the semi-parametric filter media life prediction model, uses the gradient descent method to dynamically correct the model parameters, and outputs the predicted value of the filter media life percentage for the corresponding monitoring cycle.

[0056] The early warning generation module is used to compare the predicted lifespan of the filter media and the iron and manganese removal capacity index with preset lifespan percentage threshold and capacity threshold in real time. When the iron and manganese removal capacity index is lower than the capacity threshold in a continuous monitoring period, or when the predicted lifespan percentage of the filter media is lower than the lifespan threshold, an alarm signal is generated.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] This invention continuously collects parameters such as iron and manganese ions, turbidity, and differential pressure over multiple monitoring cycles, covering multi-dimensional characteristics including differential pressure, iron and manganese removal, and backwashing events. It introduces differential pressure degradation rate and iron and manganese removal capacity indicators to quantify the system performance degradation trend and avoid interference from single instantaneous value fluctuations. Combining time-series input characteristics—capacity indicators, differential pressure rate, and backwashing events—it integrates data-driven approaches with physical laws, and is compatible with nonlinear degradation characteristics. The model parameters are updated online using the gradient descent method to adapt to the differences in filter media aging under different water quality and flow rate conditions, improving the robustness of lifespan prediction. The lifespan prediction model is dynamically corrected using backwashing markers to distinguish between normal degradation and abnormal clogging. The upper bound of the iron and manganese removal capacity indicator is required to be below a threshold for multiple consecutive cycles to avoid false alarms caused by instantaneous water quality fluctuations. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the overall method flow of the present invention;

[0060] Figure 2 This is a graph showing the relationship between the filter media lifespan factor and the proportion of predicted filter media lifespan.

[0061] Figure 3 This is a schematic diagram of the overall system flow of the present invention. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0063] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0064] Example:

[0065] Please see Figures 1-2 The present invention provides a technical solution:

[0066] A method for assessing the iron and manganese removal capacity of underground seawater, comprising the following steps:

[0067] S1: Multiple consecutive monitoring cycles are preset. Water quality parameters of seawater before and after filtration are continuously collected in each monitoring cycle. The water quality parameters include the concentration of ferrous ions and the concentration of manganese ions. The pressure difference between adjacent filtration devices and the iron and manganese turbidity of the manganese sand filtration system are collected simultaneously and continuously in each monitoring cycle. The collected data are all aggregated according to the monitoring cycle.

[0068] The multiple consecutive monitoring cycles refer to several consecutive time periods of equal length that are divided backward from the current time as the end point. Each time period is defined as a monitoring cycle. When calculating the relative rate of change of turbidity in each monitoring cycle, the continuous data of iron and manganese turbidity are integrated over the monitoring cycle and normalized by the duration of the monitoring cycle to obtain the average turbidity value in the cycle.

[0069] The continuous data of iron and manganese turbidity within the monitoring period are compared with the average turbidity value within the monitoring period. Based on the relative change between the two, the rate of change of iron and manganese turbidity in the current monitoring period relative to the average turbidity in this monitoring period is calculated and calibrated as the relative turbidity change rate of the corresponding monitoring period.

[0070] The logic for identifying turbidity abrupt changes by relative turbidity change rate is as follows: when the relative turbidity change rate reaches or exceeds a preset threshold, and the relative turbidity change rate reaches or exceeds the threshold determined by the product of the standard deviation of iron and manganese turbidity and the sensitivity coefficient within the corresponding monitoring period, when both of the above conditions are met, the event flag at that moment is set to "1", indicating that a turbidity abrupt increase event has occurred.

[0071] In the above process, if the relative rate of change of turbidity reaches or exceeds a preset threshold, and exceeds the normal range of data fluctuation within the period (defined by the product of the standard deviation and the sensitivity coefficient), it indicates an abnormal surge in turbidity. This dual-condition detection logic avoids misjudgment, distinguishing between natural fluctuations in turbidity within the period and genuine abnormal mutations, thereby improving the accuracy and reliability of identification. This method considers both the overall turbidity level of the current monitoring period and the local mutability of the data, making it suitable for monitoring the dynamic changes in iron and manganese turbidity.

[0072] The method involves calculating the relative turbidity change rate for each monitoring period based on turbidity time-series data within the monitoring cycle, and using this relative turbidity change rate to identify turbidity abrupt changes and generate backwashing operation event markers. Specifically:

[0073] In the Calculate the relative rate of change of turbidity within each monitoring period:

[0074] ;

[0075] in,

[0076] ;

[0077] Indicates the first Time points within each monitoring cycle turbidity; Indicates the first The moving average of turbidity over a monitoring period; Indicates the first Time points within each monitoring cycle The relative rate of change of turbidity; Indicates the first Initial time point within each monitoring cycle; Indicates the first The end time point within each monitoring cycle;

[0078] In the above embodiments, the changes in iron and manganese turbidity usually have a certain degree of fluctuation. Therefore, it is necessary to normalize the turbidity data within the monitoring period, i.e., calculate the average turbidity, to eliminate the influence of the overall trend within the period. Then, the local fluctuation amplitude of the turbidity data within a single period is evaluated by the relative change rate between the turbidity data and the average turbidity.

[0079] A turbidity surge event marker is generated if the following conditions are met: ,otherwise :

[0080] ;

[0081] in,

[0082] ;

[0083] The threshold representing the relative rate of change of turbidity; Indicates the first Standard deviation of turbidity within a monitoring period; Indicates the sensitivity coefficient; This indicates a sudden increase in turbidity, with 1 indicating a generated event and 0 indicating no generated event.

[0084] Calculate the number of backwashing operations within the monitoring cycle:

[0085] ;

[0086] in, Indicates the number of backwashing operations within the monitoring period; Indicates the time parameter.

[0087] In the above formula, a sudden increase in turbidity usually indicates that excessive impurities, such as iron and manganese impurities, have accumulated on the filter media surface, or that the filtration performance has declined, signifying that the filter device needs backwashing to restore normal performance. The formula determines the relative rate of change of turbidity. Does it exceed the threshold? And whether the standard deviation of turbidity significantly exceeds the sensitivity coefficient. It accurately identified the conditions that triggered backwashing; when The time indicates that a backwash event has occurred, with each marker corresponding to one backwash operation; therefore, by monitoring the time period... accumulation This method can count the number of backwashing operations that occur within the monitoring period to reflect the actual maintenance frequency and operating status of the filtration system. This calculation method directly links turbidity fluctuations with necessary backwashing operations.

[0088] S2: Based on the continuous data of iron and manganese turbidity within the monitoring period, calculate the relative change rate of turbidity in each monitoring period, identify turbidity mutations through the relative change rate of turbidity to generate backwash operation event markers, calculate the change rate of pressure difference between adjacent filter devices within the monitoring period based on the pressure difference data of adjacent filter devices within the monitoring period, and combine all the change rates to obtain the pressure difference deterioration rate.

[0089] The process involves calculating the rate of change of pressure difference between adjacent filter levels within the monitoring period based on the pressure difference data of adjacent filter levels, and then combining all the rates of change to obtain the pressure difference deterioration rate. Specifically:

[0090] In the Within each monitoring period, the rate of change of pressure difference is obtained by differentiating the pressure difference data. When the rate of change of differential pressure is positive, record the rate of differential pressure deterioration:

[0091] ;

[0092] in, Indicates the first The rate of differential pressure deterioration within each monitoring cycle; Indicates the rate of change of pressure difference; The symbol representing the rate of change of pressure difference, when When the condition is met, the value is 1; otherwise, the value is 0.

[0093] In existing technologies, a primary filter is often used to treat the influent, a secondary filter to treat particulate matter, and a high-level filter to treat iron and manganese impurities. In this embodiment, the pressure difference degradation rate collected refers to the pressure difference degradation rate between the secondary filter and the high-level filter. This pressure difference between adjacent levels is the most representative data source for the degradation rate of the filter material and is more sensitive, and can promptly reflect the clogging trend or abnormal operation of the filter in a short period of time.

[0094] The above formula uses the rate of change of pressure difference. The integral captures the cumulative change of pressure difference over time during the monitoring period. In the formula... It is a unit step function that takes a value of 1 only when the rate of change of pressure difference is positive. This ensures that the corresponding rate of change is only included in the calculation when the pressure difference increases. This is achieved by integrating the positive rate of change of pressure difference and dividing by the length of the monitoring period. The average degradation rate based on time can be obtained, which reflects the changing trend of the pressure difference between adjacent filter devices over time, thereby effectively assessing the performance degradation of the filtration system.

[0095] In the formula, the unit step function is used. To determine the rate of change of pressure difference The positive and negative cases; when When the pressure difference is increasing, the system records the rate of change; when When the pressure difference does not increase or decrease, the system ignores the rate of change; at the time indicated by ... Within each monitoring cycle, for all time points Positive pressure differential change rate Integrate to obtain the cumulative deterioration of the pressure differential over time; divide this cumulative value by the length of the monitoring period. Calculate the rate of pressure differential degradation.

[0096] S3: Calculate the difference between the influent and effluent concentrations of ferrous ions and manganese ions in the monitoring period to obtain the iron removal amount and manganese removal amount in the monitoring period. After obtaining the iron removal amount and manganese removal amount in each monitoring period, use them as input to the preset multi-factor iron and manganese removal capacity calculation model and output the iron and manganese removal capacity index of the corresponding monitoring period.

[0097] The difference between the influent and effluent water concentrations of ferrous ions and manganese ions during the monitoring period is calculated to obtain the iron and manganese removal amounts for that period. Specifically:

[0098] The differences in influent and effluent concentrations of ferrous and manganese ions within the monitoring period are calculated separately. The iron and manganese removal amounts at each moment in the monitoring period are integrated to obtain the total iron and manganese removal amounts for the entire monitoring period.

[0099] ;

[0100] ;

[0101] in, Indicates the first Iron removal amount within a monitoring cycle; Indicates the first Manganese removal amount within a monitoring cycle; Indicates the first Within each monitoring period Iron removal rate at any time; Indicates the first Within each monitoring period Manganese removal rate at any given time.

[0102] After obtaining the iron and manganese removal amounts for each monitoring period, these amounts are used as inputs to a preset multi-factor iron and manganese removal capacity calculation model, which outputs the iron and manganese removal capacity index for the corresponding monitoring period. The specific steps are as follows:

[0103] Based on iron and manganese removal rates, a multi-factor removal capacity calculation model is constructed:

[0104] ;

[0105] in, Indicates the first The removal capacity index values ​​of iron and manganese within each monitoring cycle; Indicates the maximum iron removal capacity designed; Indicate the maximum manganese removal capacity designed; This represents the iron removal weighting coefficient; This represents the weighting coefficient for manganese removal, where , ,and .

[0106] In the above formula, the amount of iron and manganese removed... and It is real data reflecting removal capacity. As a comprehensive indicator, the contributions of both must be considered simultaneously, so the formula combines them in a weighted manner; the denominator provides a fixed reference standard that reflects whether the removal capacity of the current cycle has achieved the design target. The functional form avoids the problems of being too large or too small that may occur when directly using a linear ratio. If the actual removal amount is much smaller than the designed removal amount, The value will not infinitely approach 0 if the actual removal amount is much greater than the designed removal amount. The growth of [value] will gradually level off, avoiding the impact of excessively large values ​​on the results; through weighting and The settings can be adjusted to modify the iron and manganese removal capacity according to actual needs. Contributions;

[0107] because , ,and This means that these two weighting coefficients should take values ​​between 0 and 1, and their sum should be 1, thus ensuring that they together form a complete relative weight allocation. This setting makes the removal capability index... It can reflect the relative contributions of iron and manganese based on a comprehensive consideration of their removal effects, facilitating a fair comparison and evaluation of removal capacity, while maintaining the linearity of the model and ensuring the applicability of the formula.

[0108] in, This indicates the removal capacity index value of iron and manganese within the detection period. The closer the value is to 0, the closer the actual removal capacity of the system is to the design limit, and the higher the risk of filter media saturation or failure; the weighted value of the actual removal amount directly affects The higher its proportion, The higher the value, the stronger the current removal ability; by adjusting... and This can specifically amplify the monitoring sensitivity of iron or manganese pollution. , and Positive correlation; as the actual removal amount increases, the molecular... Enlargement, leading to The value increased; , As the benchmark value in the denominator, its increase will decrease the ratio, indirectly suppressing... Growth; through The function restricts the output to a monotonically increasing curve to avoid extreme values ​​affecting model stability;

[0109] S4: The iron and manganese removal capacity index, backwashing operation event marker, and differential pressure deterioration rate of multiple consecutive monitoring cycles are used as time-series input features and input into the semi-parametric filter media life prediction model. The model outputs the predicted life percentage of the filter media in the corresponding monitoring cycle and uses the gradient descent method to dynamically correct the model parameters.

[0110] The process of using iron and manganese removal capacity indicators, backwashing operation event markers, and differential pressure deterioration rates from multiple consecutive monitoring periods as time-series input features, inputting them into a semi-parametric filter media life prediction model, and outputting the predicted lifespan percentage of the filter media for the corresponding monitoring period is as follows:

[0111] Constructing the filter media lifespan factor:

[0112] ;

[0113] in, Indicates the first Filter media lifespan factor within each monitoring cycle; This represents the theoretical optimal value for the number of backwashes per unit time. This represents the theoretically optimal value for the rate of pressure differential degradation. This represents the theoretical optimal value of the iron and manganese removal capacity index; This indicates the sensitivity coefficient to the impact of the number of backwashing cycles on the lifespan. This represents the sensitivity coefficient to the impact of differential pressure degradation rate on lifetime. This indicates the sensitivity coefficient to the impact of iron and manganese removal capacity on the product's lifespan. Indicates the first Number of backwashing operations within a monitoring cycle;

[0114] In the above formula, Weakening by denominator terms , Indicates when Deviation When the square term increases rapidly, it weakens the square term through the denominator. The value indicates Deviations result in a performance decrease, but the change is relatively gradual. Significantly reduced through the exponential term , show Deviation from optimal value When, dependent variable It will drop rapidly, indicating It is a variable that is quite sensitive to the impact on performance; By considering several terms show Will be Having a positive effect indicates that the variable will have a certain positive impact after deviating from the optimal value, but the impact gradually decreases.

[0115] in, This represents the filter media life factor, used to quantify the frequency of backwashing operations. Differential pressure deterioration rate and iron and manganese removal capacity The overall impact on the remaining lifespan of the filter media; the lower the output value, the more severe the loss of filter media lifespan, providing a quantitative basis for triggering cleaning or replacement; Deviation from theoretical optimal value The quadratic term in the denominator increases, leading to A decrease reflects either excessive cleaning leading to wasted energy or insufficient cleaning. Deviation from theoretical optimal value The exponential decay intensified, leading to Significantly reduced, indicating the degree of filter clogging; Deviation from theoretical optimal value Increase in several outputs, improvement The value indicates the risk of filter media adsorption saturation or failure; through secondary terms... A nonlinear penalty is introduced, which increases the denominator when the value deviates from the optimal value, whether too high or too low, to suppress [the deviation from the optimal value]. ; Index term The dominant negative decay occurs when the rate of degradation exceeds the optimal value, leading to a sharp decrease in the exponential value and accelerating lifespan loss; the logarithmic term... It provides positive gain, and when the actual capability exceeds the optimal value, it can partially offset the losses caused by other factors.

[0116] in, The impact of backwashing frequency on filter media wear and performance recovery should be evaluated, and the degree of its significant impact on filter media life is usually determined by analyzing historical data. Then it is necessary to consider the rate of pressure differential degradation and the rate of filter media performance decline. Usually, it is necessary to analyze the relationship between pressure differential change and filter media life through experimental data to set the parameters. The impact of iron and manganese removal capacity on the working efficiency and service life of filter media should be quantitatively evaluated.

[0117] Using the filter media lifespan factor as a semi-parametric adjustment factor, a semi-parametric model is constructed to predict the proportion of filter media lifespan:

[0118] ;

[0119] in, Indicates the predicted first The percentage of remaining lifespan of the filter media in each monitoring cycle.

[0120] In the above formula, parameter 50 is the inflection point of the function; 0.1 represents the sensitivity coefficient; The output of this value is always between 0 and 1, and can be used to represent percentages; where, This represents the percentage of the predicted remaining lifespan of the filter media, and is a quantified percentage of the remaining lifespan of the filter media, expressed using the sigmoid function. Nonlinear compression is used to avoid lifetime jumps and enhance the robustness of prediction results; when At that time, remaining lifespan Rapidly decaying to a low value triggers a replacement warning; As a semi-parametric adjustment factor, it integrates dynamic parameters such as backwashing frequency, differential pressure deterioration rate, and iron and manganese removal capacity to characterize the overall wear and tear of filter media. A higher value indicates less filter media wear and longer remaining lifespan. Approaching 100%; when At that time, the exponent term Rapid growth led to A sharp decline enhances sensitivity to severely deteriorated conditions; and It shows a monotonically positive correlation. With each increase, the remaining lifespan Simultaneous improvement;

[0121] In this embodiment, 10 are selected. By using the filter media lifespan factor as a semi-parametric adjustment factor, a semi-parametric model was constructed to predict the proportion of filter media lifespan. The experimental data are shown in Table 1.

[0122] Table 1: Relationship between Filter Media Lifetime Factor and Filter Media Lifetime Ratio

[0123]

[0124] As can be seen from the table above, when When the function approaches 1, Approaching 100%, the filter media is in ideal condition; when The function approaches 0. When the filter media approaches 0%, it becomes completely ineffective. A higher value indicates less filter media wear and a longer remaining lifespan. Approaching 100%;

[0125] The gradient descent method is used to dynamically adjust the model parameters, specifically as follows:

[0126] Mean square error of the filter media lifespan influencing factors in the semi-parametric model:

[0127] ;

[0128] in, This represents the theoretical filter media lifespan factor value. The mean square error representing the loss of filter media influence factors;

[0129] Calculate separately for , , gradient:

[0130] ;

[0131] ;

[0132] ;

[0133] in, express for The gradient; express for The gradient; express for The gradient;

[0134] Parameter updates are performed using gradient descent.

[0135] ;

[0136] in, This represents the parameters before the update, where To affect the sensitivity coefficient index, , representing the sensitivity coefficient of the number of backwashing cycles to the lifespan. , representing the sensitivity coefficient to the impact of differential pressure degradation rate on lifetime. , representing the sensitivity coefficient of the iron and manganese removal capacity index to the lifespan; Indicates the updated parameters ; Indicates the learning rate; The loss function represents the... The gradient.

[0137] S5: Based on the predicted lifespan of the filter media and the iron and manganese removal capacity index, compare them in real time with the preset lifespan percentage threshold and capacity threshold respectively. When the iron and manganese removal capacity index is lower than the capacity threshold in a continuous monitoring period, or the predicted lifespan percentage of the filter media is lower than the lifespan threshold, an alarm signal is generated.

[0138] Set the remaining lifespan threshold of the filter media as follows: Determine and calibrate the predicted remaining lifespan of the filter media for the most recent monitoring period. ,like Then, trace back... If the predicted remaining life of the filter media is less than [a certain value] for all monitoring cycles traced back, then [the following is a separate, unrelated sentence:] ...for each monitoring cycle, if the predicted remaining life of the filter media for all monitoring cycles traced back is less than [a certain value]... If the filter media life is insufficient, a warning signal will be issued; otherwise, no warning signal will be issued.

[0139] Set the capability threshold as Without issuing an early warning signal, determine and calibrate the iron and manganese removal capacity index of the monitoring period closest to the current time. ,like Then trace back... If the iron and manganese removal capacity index for all monitored periods is less than [a certain value], then [the following is a possible interpretation:] ... If so, an abnormal warning signal for iron and manganese removal capacity will be issued.

[0140] In the above process, tracing back... Each monitoring cycle refers to a continuous backward regression from the most recent monitoring cycle at the current moment. Each cycle includes, in sequence: , until These monitoring cycles are typically divided into fixed time intervals, such as daily, weekly, or monthly, and are tracked back through these cycles. A continuous historical cycle can be used to assess the continuous trend of the remaining life prediction of the filter media or the iron and manganese removal capacity index over a period of time, thereby determining whether the problem is long-term or sporadic and providing a basis for issuing early warning signals.

[0141] The present invention also provides a system for assessing the iron and manganese removal capacity of underground seawater, the system being used to perform the above-described assessment method, comprising:

[0142] The data acquisition module is used to preset multiple continuous monitoring cycles. In each monitoring cycle, it continuously collects water quality index parameters of seawater before and after filtration. The water quality index parameters include the concentration of ferrous ions and the concentration of manganese ions. Simultaneously, it continuously collects the pressure difference between adjacent levels of filtration devices and the iron and manganese turbidity of the manganese sand filtration system in each monitoring cycle, and collects the data according to the monitoring cycle.

[0143] The turbidity identification and differential pressure degradation calculation module is used to calculate the relative turbidity change rate of each monitoring period based on the continuous data of iron and manganese turbidity within the monitoring period. It generates backwash operation event markers by identifying turbidity abrupt changes through the relative turbidity change rate. Based on the pressure difference data of adjacent level filter devices within the monitoring period, it calculates the change rate of pressure difference between adjacent level filter devices within the monitoring period and integrates all change rates to obtain the differential pressure degradation rate.

[0144] The iron and manganese removal calculation module is used to calculate the difference between the influent and effluent concentrations of ferrous ions and manganese ions in the monitoring period, respectively, to obtain the iron and manganese removal amount in that monitoring period. After obtaining the iron and manganese removal amount in each monitoring period, it is used as input to be sent to the preset multi-factor iron and manganese removal capacity calculation model, and the iron and manganese removal capacity index of the corresponding monitoring period is output.

[0145] The filter media life prediction module takes the iron and manganese removal capacity index, backwash operation event markers, and differential pressure deterioration rate of multiple consecutive monitoring cycles as time-series input features, inputs them into the semi-parametric filter media life prediction model, uses the gradient descent method to dynamically correct the model parameters, and outputs the predicted value of the filter media life percentage for the corresponding monitoring cycle.

[0146] The early warning generation module is used to compare the predicted lifespan of the filter media and the iron and manganese removal capacity index with preset lifespan percentage threshold and capacity threshold in real time. When the iron and manganese removal capacity index is lower than the capacity threshold in a continuous monitoring period, or when the predicted lifespan percentage of the filter media is lower than the lifespan threshold, an alarm signal is generated.

[0147] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0148] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0149] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0150] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for evaluating the iron and manganese removal capacity of underground seawater, characterized by the following steps: include: S1: Multiple consecutive monitoring cycles are preset. Water quality parameters of seawater before and after filtration are continuously collected in each monitoring cycle. The water quality parameters include the concentration of ferrous ions and the concentration of manganese ions. The pressure difference between adjacent filtration devices and the iron and manganese turbidity of the manganese sand filtration system are collected simultaneously and continuously in each monitoring cycle. The collected data are all aggregated according to the monitoring cycle. S2: Based on the continuous data of iron and manganese turbidity within the monitoring period, calculate the relative change rate of turbidity in each monitoring period, identify turbidity mutations through the relative change rate of turbidity to generate backwash operation event markers, calculate the change rate of pressure difference between adjacent filter devices within the monitoring period based on the pressure difference data of adjacent filter devices within the monitoring period, and combine all the change rates to obtain the pressure difference deterioration rate. S3: Calculate the difference between the influent and effluent concentrations of ferrous ions and manganese ions in the monitoring period to obtain the iron removal amount and manganese removal amount in the monitoring period. After obtaining the iron removal amount and manganese removal amount in each monitoring period, use them as input to the preset multi-factor iron and manganese removal capacity calculation model and output the iron and manganese removal capacity index of the corresponding monitoring period. S4: The iron and manganese removal capacity index, backwashing operation event marker, and differential pressure deterioration rate of multiple consecutive monitoring cycles are used as time-series input features and input into the semi-parametric filter media life prediction model. The model outputs the predicted life percentage of the filter media in the corresponding monitoring cycle and uses the gradient descent method to dynamically correct the model parameters. S5: Based on the predicted lifespan of the filter media and the iron and manganese removal capacity index, compare them in real time with the preset lifespan percentage threshold and capacity threshold respectively. When the iron and manganese removal capacity index is lower than the capacity threshold in a continuous monitoring period, or the predicted lifespan percentage of the filter media is lower than the lifespan threshold, generate an alarm signal. The process of using iron and manganese removal capacity indicators, backwashing operation event markers, and differential pressure deterioration rates from multiple consecutive monitoring periods as time-series input features, inputting them into a semi-parametric filter media life prediction model, and outputting the predicted lifespan percentage of the filter media for the corresponding monitoring period is as follows: Constructing the filter media lifespan factor: in, X represents the filter media lifespan factor within the i-th monitoring cycle; 1,opt This represents the theoretical optimal value for the number of backwashes per unit time; X 2,opt X represents the theoretically optimal rate of pressure differential degradation; 3,opt k1 represents the theoretical optimal value of the iron and manganese removal capacity index; k2 represents the sensitivity coefficient of the impact of the number of backwashings within the monitoring cycle on the service life; k3 represents the sensitivity coefficient of the impact of the iron and manganese removal capacity index on the service life. This represents the number of backwashing operations within the i-th monitoring period; Using the filter media lifespan factor as a semi-parametric adjustment factor, a semi-parametric model is constructed to predict the proportion of filter media lifespan: Among them, L i This represents the percentage of the remaining lifespan of the filter media in the predicted i-th monitoring cycle.

2. The method for evaluating the iron and manganese removal capacity of underground seawater according to claim 1, characterized in that, The multiple consecutive monitoring cycles refer to several consecutive time periods of equal length that are divided backward from the current time as the end point. Each time period is defined as a monitoring cycle. When calculating the relative rate of change of turbidity in each monitoring cycle, the continuous data of iron and manganese turbidity are integrated over the monitoring cycle and normalized by the duration of the monitoring cycle to obtain the average turbidity value in the cycle. The continuous data of iron and manganese turbidity within the monitoring period are compared with the average turbidity value within the monitoring period. Based on the relative change between the two, the rate of change of iron and manganese turbidity in the current monitoring period relative to the average turbidity in this monitoring period is calculated and calibrated as the relative turbidity change rate of the corresponding monitoring period. The logic for identifying turbidity abrupt changes by the relative rate of change of turbidity is as follows: when the relative rate of change of turbidity reaches or exceeds a preset threshold, and the relative rate of change of turbidity reaches or exceeds the threshold determined by the product of the standard deviation of iron and manganese turbidity and the sensitivity coefficient within the corresponding monitoring period, when both of the above conditions are met, the event flag at that moment is set to "1", indicating that a turbidity abrupt increase event has occurred.

3. The method for evaluating the iron and manganese removal capacity of underground seawater according to claim 1, characterized in that, The process involves calculating the rate of change of pressure difference between adjacent filter levels within the monitoring period based on the pressure difference data of adjacent filter levels, and then combining all the rates of change to obtain the pressure difference deterioration rate. Specifically: During the i-th monitoring period, the rate of change of pressure difference P is obtained by differentiating the pressure difference data. i (t), when the rate of change of pressure difference is positive, record the rate of pressure difference deterioration: in, P represents the rate of differential pressure degradation during the i-th monitoring period; i (t) represents the rate of change of pressure difference; H[P i [(t)] represents the symbol for the rate of change of pressure difference. It takes the value 1 when P1(t) > 0, and 0 otherwise; t 1,i t represents the start time within the i-th monitoring period; 2,i This represents the end time within the i-th monitoring period.

4. The method for evaluating the iron and manganese removal capacity of underground seawater according to claim 3, characterized in that, The difference between the influent and effluent water concentrations of ferrous ions and manganese ions during the monitoring period is calculated to obtain the iron and manganese removal amounts for that period. Specifically: The differences in influent and effluent concentrations of ferrous and manganese ions within the monitoring period are calculated separately. The iron and manganese removal amounts at each moment in the monitoring period are integrated to obtain the total iron and manganese removal amounts for the entire monitoring period. in, This represents the amount of iron removed during the i-th monitoring period; This represents the amount of manganese removed during the i-th monitoring period; This represents the amount of iron removed at time t within the i-th monitoring period; This represents the amount of manganese removed at time t within the i-th monitoring period.

5. The method for evaluating the iron and manganese removal capacity of underground seawater according to claim 4, characterized in that, After obtaining the iron and manganese removal amounts for each monitoring period, these amounts are used as inputs to a preset multi-factor iron and manganese removal capacity calculation model, which outputs the iron and manganese removal capacity index for the corresponding monitoring period. The specific steps are as follows: Based on iron and manganese removal rates, a multi-factor removal capacity calculation model is constructed: in, This represents the iron and manganese removal capacity index value within the i-th monitoring period; Indicates the maximum iron removal capacity designed; The maximum manganese removal capacity is shown in the design; ω1 represents the iron removal weight coefficient; ω2 represents the manganese removal weight coefficient, where ω1>0, ω2>0, and ω1+ω2=1.

6. The method for evaluating the iron and manganese removal capacity of underground seawater according to claim 1, characterized in that, The gradient descent method is used to dynamically adjust the model parameters, specifically as follows: Mean square error of the filter media lifespan influencing factors in the semi-parametric model: Among them, L new This represents the theoretical filter media lifespan factor value. The mean square error representing the loss of filter media influence factors; Calculate separately For the gradients of k1, k2, and k3: in, express The gradient of k1; express The gradient of k2; express The gradient of k3; Parameter updates are performed using gradient descent. Where, k w The parameters before the update are represented, where w is the sensitivity coefficient index. w = 1 represents the sensitivity coefficient of the number of backwashes within the monitoring period on the lifespan, w = 2 represents the sensitivity coefficient of the pressure differential deterioration rate on the lifespan, and w = 3 represents the sensitivity coefficient of the iron and manganese removal capacity index on the lifespan. This indicates the updated parameter k. w η represents the learning rate; The loss function represents the loss function for k. w The gradient.

7. The method for evaluating the iron and manganese removal capacity of underground seawater according to claim 1, characterized in that, The specific steps for generating the alarm signal are as follows: Set the remaining lifespan threshold of the filter media to L. threshold Determine the predicted remaining lifespan of the filter media for the most recent monitoring period and label it as L. now If L now <L threshold If the predicted remaining lifespan of the filter media is less than L for all the preceding monitoring cycles, then the process continues for the next M monitoring cycles. threshold If the filter media life is insufficient, a warning signal will be issued; otherwise, no warning signal will be issued. Set the capability threshold to C threshold Without issuing an early warning signal, determine and label the iron and manganese removal capacity index of the monitoring period closest to the current time as M. now If M now <C threshold Then, trace back M monitoring periods. If the iron and manganese removal capacity index for all traced monitoring periods is less than C... threshold If so, an abnormal warning signal for iron and manganese removal capacity will be issued.

8. A system for evaluating the iron and manganese removal capacity of underground seawater, characterized in that: The evaluation system is used to perform the evaluation method according to any one of claims 1-7, including: The data acquisition module is used to preset multiple continuous monitoring cycles. In each monitoring cycle, it continuously collects water quality index parameters of seawater before and after filtration. The water quality index parameters include the concentration of ferrous ions and the concentration of manganese ions. Simultaneously, it continuously collects the pressure difference between adjacent levels of filtration devices and the iron and manganese turbidity of the manganese sand filtration system in each monitoring cycle, and collects the data according to the monitoring cycle. The turbidity identification and differential pressure degradation calculation module is used to calculate the relative turbidity change rate of each monitoring period based on the continuous data of iron and manganese turbidity within the monitoring period. It generates backwash operation event markers by identifying turbidity abrupt changes through the relative turbidity change rate. Based on the pressure difference data of adjacent level filter devices within the monitoring period, it calculates the change rate of pressure difference between adjacent level filter devices within the monitoring period and integrates all change rates to obtain the differential pressure degradation rate. The iron and manganese removal calculation module is used to calculate the difference between the influent and effluent concentrations of ferrous ions and manganese ions in the monitoring period, respectively, to obtain the iron and manganese removal amount in that monitoring period. After obtaining the iron and manganese removal amount in each monitoring period, it is used as input to be sent to the preset multi-factor iron and manganese removal capacity calculation model, and the iron and manganese removal capacity index of the corresponding monitoring period is output. The filter media life prediction module takes the iron and manganese removal capacity index, backwash operation event markers, and differential pressure deterioration rate of multiple consecutive monitoring cycles as time-series input features, inputs them into the semi-parametric filter media life prediction model, uses the gradient descent method to dynamically correct the model parameters, and outputs the predicted value of the filter media life percentage for the corresponding monitoring cycle. The early warning generation module is used to compare the predicted filter media lifespan and the iron and manganese removal capacity index with preset lifespan percentage thresholds and capacity thresholds in real time. When the iron and manganese removal capacity index falls below the capacity threshold for a continuous monitoring period, or An alarm signal is generated when the predicted percentage of filter media lifespan is lower than the lifespan threshold.

Images