Online detection method and system for the impurity removal efficiency of fluid iron separator

Abstract

This invention relates to the field of fluid impurity removal detection technology, and discloses an online detection method and system for the impurity removal efficiency of a fluid iron separator. The method includes: measuring the instantaneous content of magnetic impurities at the inlet and outlet in real time using a high-frequency electromagnetic induction sensor; monitoring the change in magnetic flux on the surface of the magnetic medium in real time using a miniature fluxgate sensor array; calculating the apparent impurity removal efficiency based on the impurity content at the inlet and outlet using an edge computing gateway, and simultaneously calculating the magnetic medium adsorption saturation based on the magnetic flux decay curve; weighting the apparent impurity removal efficiency using the magnetic medium adsorption saturation as a weight for adsorption effectiveness; when the magnetic medium adsorption saturation is lower than a preset saturation threshold, attenuation is performed according to the ratio of the apparent impurity removal efficiency multiplied by the magnetic medium adsorption saturation to the preset saturation threshold to obtain the true impurity removal efficiency. This invention solves the problem in existing technologies that cannot distinguish whether impurities are effectively retained or merely pass through the iron separator, achieving accurate detection of the true impurity removal efficiency.

Description

Technical Field

[0001] This invention relates to the field of fluid impurity removal detection technology, and in particular to a method and system for online detection of the impurity removal efficiency of fluid iron separators. Background Technology

[0002] Fluid separators are widely used in coal, power, and chemical industries to remove ferromagnetic impurities from fluids, protecting downstream equipment and ensuring product quality. Current online detection methods for impurity removal efficiency typically involve deploying electromagnetic induction sensors at the inlet and outlet pipes, calculating the removal efficiency by measuring the difference between the impurity content at the inlet and outlet. This method is simple to operate and has a fast response time, making it the mainstream online detection method in industrial settings.

[0003] However, existing detection methods have a fundamental flaw: they can only detect whether impurities have passed through the magnetic separator, but cannot determine whether the impurities have been effectively retained. When the magnetic medium's adsorption capacity decreases or approaches saturation, some impurities may only briefly pass through the magnetic medium area without being effectively captured. However, because the outlet detection value is still low, the system may misjudge the impurity removal efficiency as normal. This confusion between apparent removal and true retention leads to detection results that deviate significantly from reality. Operators cannot promptly ascertain the true operating status of the magnetic separator, and misjudgments can easily delay cleaning or replacement, ultimately causing damage to downstream equipment or product quality incidents.

[0004] Therefore, this invention proposes an online detection method and system for the impurity removal efficiency of fluid iron separators. Summary of the Invention

[0005] This invention provides an online detection method and system for the impurity removal efficiency of fluid separators. By introducing magnetic medium adsorption saturation as a correction weight on the basis of traditional inlet and outlet detection, the detection blind zone between apparent removal and actual retention is eliminated. This upgrades the impurity removal efficiency detection result from reflecting whether impurities have passed through the equipment to accurately characterizing whether impurities have been effectively retained, providing a true and reliable basis for separator condition assessment and maintenance decisions.

[0006] This invention provides an online detection method for the impurity removal efficiency of a fluid iron separator, comprising: High-frequency electromagnetic induction sensors are deployed in the inlet and outlet pipes of the fluid separator to measure the instantaneous content of magnetic impurities in the fluid in real time, thereby obtaining the inlet impurity content sequence and the outlet impurity content sequence. A miniature fluxgate sensor array is deployed on the surface of the magnetic medium inside the fluid separator. The change in magnetic flux on the surface of the magnetic medium is monitored in real time by the miniature fluxgate sensor array, and the magnetic flux decay curve is obtained. The edge computing gateway receives the inlet impurity content sequence, the outlet impurity content sequence, and the magnetic flux attenuation curve. The edge computing gateway calculates the apparent impurity removal efficiency based on the inlet impurity content sequence and the outlet impurity content sequence, and calculates the magnetic medium adsorption saturation based on the magnetic flux decay curve. The edge computing gateway uses the magnetic medium adsorption saturation as the weight of adsorption effectiveness to perform weighted fusion of apparent impurity removal efficiency. When the magnetic medium adsorption saturation is lower than the preset saturation threshold, the confidence of apparent impurity removal efficiency is reduced and the apparent impurity removal efficiency is attenuated according to the ratio of the actual impurity removal efficiency to the apparent impurity removal efficiency multiplied by the magnetic medium adsorption saturation to the preset saturation threshold, so as to obtain the actual impurity removal efficiency.

[0007] Furthermore, the steps for calculating the magnetic medium adsorption saturation based on the magnetic flux decay curve include: The edge computing gateway performs a first-order difference operation on the magnetic flux decay curve to obtain the magnetic flux decay rate curve. The edge computing gateway performs a second-order difference operation on the magnetic flux decay rate curve to obtain the magnetic flux decay acceleration curve. The edge computing gateway searches for the moment when the magnetic flux decay rate approaches a preset minimum rate threshold in the magnetic flux decay acceleration curve, which is the turning point when the magnetic medium surface adsorption layer enters the critical state of collapse from the stable adsorption state. The edge computing gateway uses the magnetic flux value at the turning point as the initial saturation magnetic flux and the ratio of the current magnetic flux to the initial saturation magnetic flux as the magnetic medium adsorption saturation.

[0008] Furthermore, the steps for calculating the magnetic medium adsorption saturation based on the magnetic flux decay curve include: The edge computing gateway acquires the magnetic flux values ​​of multiple sampling points within a preset historical time period for the magnetic flux decay curve; The edge computing gateway performs exponential function fitting on the magnetic flux values ​​of multiple sampling points to obtain the magnetic flux decay fitting function; The edge computing gateway extracts the decay time constant of the magnetic flux decay fitting function, and uses the ratio of the decay time constant to the preset reference time constant as the magnetic medium adsorption saturation. The preset reference time constant is obtained by calibrating the adsorption experiment on a brand-new fluid iron separator of the same specification under standard working conditions.

[0009] Furthermore, it also includes steps for correcting the uniformity of the entire flow field: Distributed magnetic sensors are deployed at the inlet section, the middle section, and the outlet section of the internal flow channel of the fluid separator. The magnetic impurity concentration data at different spatial locations are collected synchronously through the distributed magnetic sensors. The edge computing gateway receives magnetic impurity concentration data collected by the distributed magnetic sensor and calculates the axial concentration gradient based on the magnetic impurity concentration data of the inlet section, the middle section, and the outlet section. The edge computing gateway incorporates the axial concentration gradient as a flow field uniformity coefficient into the calculation of the actual impurity removal efficiency to obtain the final actual impurity removal efficiency.

[0010] Furthermore, the steps of incorporating the axial concentration gradient as a flow field uniformity coefficient into the calculation of the actual impurity removal efficiency include: The edge computing gateway acquires magnetic impurity concentration data for fine particle size range, medium particle size range, and coarse particle size range collected by distributed magnetic sensors. The edge computing gateway calculates the axial concentration gradients for the fine particle size range, the medium particle size range, and the coarse particle size range, respectively. The edge computing gateway performs a weighted summation of the axial concentration gradients of the fine particle size range, the medium particle size range, and the coarse particle size range according to preset weights to obtain the particle size weighted uniformity coefficient. When the particle size weighted uniformity coefficient exceeds the preset uniformity threshold, the edge computing gateway performs a second correction on the actual impurity removal efficiency. Finally, the actual impurity removal efficiency is equal to the actual impurity removal efficiency multiplied by the ratio of the preset uniformity threshold to the particle size weighted uniformity coefficient.

[0011] Furthermore, a dynamic compensation step is included before the step of calculating the apparent impurity removal efficiency based on the import impurity content sequence and the export impurity content sequence: A flow velocity sensor and a viscosity sensor are deployed in the inlet pipe of the fluid separator. The flow velocity sensor collects the fluid flow velocity in real time, and the viscosity sensor collects the fluid viscosity in real time. The edge computing gateway receives fluid velocity and fluid viscosity, and determines the dynamic compensation coefficient according to the ratio of fluid velocity to reference velocity and the ratio of fluid viscosity to reference viscosity in a linear function relationship. The dynamic compensation coefficient is equal to 1 plus the difference between the velocity compensation coefficient and the ratio of fluid velocity to reference velocity minus 1, plus the difference between the viscosity compensation coefficient and the ratio of fluid viscosity to reference viscosity minus 1. The edge computing gateway corrects the instantaneous magnetic impurity content in the inlet impurity content sequence and the instantaneous magnetic impurity content in the outlet impurity content sequence according to the dynamic compensation coefficient. The corrected instantaneous magnetic impurity content is equal to the original instantaneous magnetic impurity content multiplied by the dynamic compensation coefficient. The edge computing gateway uses the corrected inlet impurity content sequence and the corrected outlet impurity content sequence as input data to calculate the apparent impurity removal efficiency.

[0012] Furthermore, the steps of obtaining the import impurity content sequence and the export impurity content sequence include a sub-step of multi-frequency electromagnetic response identification: The high-frequency electromagnetic induction sensor alternately emits electromagnetic waves at a first detection frequency and a second detection frequency, wherein the first detection frequency is lower than the second detection frequency. The edge computing gateway receives a first response signal corresponding to a first detection frequency and a second response signal corresponding to a second detection frequency. According to the skin effect principle, when the first response signal is greater than or equal to the second response signal, the edge computing gateway takes the difference between the first response signal and the second response signal as the contribution value of weak magnetic impurities, and takes the remaining part of the second response signal after subtracting the contribution value of weak magnetic impurities as the content value of strong magnetic ferromagnetic impurities. When the first response signal is less than the second response signal, the contribution value of the weak magnetic impurities is set to zero, and the second response signal is directly used as the content value of the strong magnetic ferromagnetic impurities. The edge computing gateway regenerates the inlet and outlet impurity content sequences based on the content values ​​of strongly magnetic and ferromagnetic impurities, so that each content value in the inlet and outlet impurity content sequences corresponds to a strongly magnetic and ferromagnetic impurity that can be effectively removed by the fluid iron separator.

[0013] Furthermore, after obtaining the true impurity removal efficiency, a trend prediction step is also included: The edge computing gateway obtains the actual noise removal efficiency calculated at multiple time points within a historical period, forming a historical sequence of actual noise removal efficiency. The edge computing gateway performs time-series analysis on the historical sequence of the actual noise removal efficiency to extract the decay rate of the actual noise removal efficiency; The edge computing gateway predicts the trend of real noise removal efficiency in the future based on the decay rate of real noise removal efficiency. When the predicted real noise removal efficiency will drop below the target efficiency threshold within a preset time window, an efficiency decay warning signal is generated.

[0014] Furthermore, after obtaining the actual impurity removal efficiency, a step to indicate the cleaning effect is also included: The edge computing gateway obtains the historical cleaning records of the fluid iron separator. The historical cleaning records include the magnetic media adsorption saturation before each cleaning operation, the magnetic media adsorption saturation after each cleaning operation, and the actual impurity removal efficiency recovery value after each cleaning operation. The edge computing gateway establishes a cleaning effect prediction model based on historical cleaning records. The cleaning effect prediction model is used to characterize the relationship between the magnetic medium adsorption saturation and the actual cleanliness recovery value after cleaning. The edge computing gateway inputs the current magnetic medium adsorption saturation into the cleaning effect prediction model to obtain the predicted efficiency recovery value after the current cleaning operation, and generates cleaning timing suggestions and expected effect prompts based on the predicted efficiency recovery value, without correcting the current actual impurity removal efficiency.

[0015] This invention provides an online detection system for the impurity removal efficiency of a fluid iron separator, comprising: A high-frequency electromagnetic induction sensor is deployed in the inlet and outlet pipes of a fluid separator to measure the instantaneous content of magnetic impurities in the fluid in real time, generating inlet and outlet impurity content sequences. A miniature fluxgate sensor array is deployed on the surface of the magnetic medium inside the fluid separator to monitor the change of magnetic flux on the surface of the magnetic medium in real time and generate a magnetic flux decay curve. The edge computing gateway is connected to a high-frequency electromagnetic induction sensor and a miniature fluxgate sensor array, respectively, to receive the inlet impurity content sequence, the outlet impurity content sequence, and the flux decay curve. The edge computing gateway includes an apparent cleanup efficiency calculation unit, a magnetic medium adsorption saturation calculation unit, and a true cleanup efficiency fusion unit; The apparent impurity removal efficiency calculation unit is used to calculate the apparent impurity removal efficiency based on the inlet impurity content sequence and the outlet impurity content sequence. The magnetic medium adsorption saturation calculation unit is used to calculate the magnetic medium adsorption saturation based on the magnetic flux decay curve. The true impurity removal efficiency fusion unit is used to weight and fuse the apparent impurity removal efficiency by using the magnetic medium adsorption saturation as the adsorption effectiveness weight. When the magnetic medium adsorption saturation is lower than the preset saturation threshold, the true impurity removal efficiency fusion unit reduces the confidence of the apparent impurity removal efficiency and attenuates the apparent impurity removal efficiency according to the ratio of the true impurity removal efficiency to the apparent impurity removal efficiency multiplied by the magnetic medium adsorption saturation to the preset saturation threshold, so as to obtain the true impurity removal efficiency.

[0016] The beneficial effects of this invention compared to existing technologies are as follows: it overcomes the shortcomings of existing detection technologies that only focus on the difference in impurity content between inlets and outlets and cannot perceive the adsorption state of the magnetic medium itself. This method, by introducing the magnetic medium adsorption saturation as a core monitoring dimension, can identify in real time the state where the magnetic medium's adsorption capacity is decreasing or approaching saturation. When the adsorption saturation is below a threshold, the apparent impurity removal efficiency is attenuated and corrected, thereby removing penetrating impurities that are mistakenly judged as removed in traditional detection methods from the efficiency calculation, ensuring that the detection results truly reflect the actual retention of impurities. Operators can accurately judge the working status of the iron separator based on the corrected true impurity removal efficiency, and take timely cleaning or replacement measures when the adsorption capacity decreases, avoiding damage to downstream equipment or product quality accidents due to misjudgment, and significantly improving the operational reliability and process control accuracy of the fluid iron removal system.

[0017] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall flowchart of the online detection method for the impurity removal efficiency of a fluid iron separator in this embodiment of the invention; Figure 2 This is a flowchart of the method for calculating the adsorption saturation of magnetic media based on the magnetic flux decay curve in an embodiment of the present invention. Detailed Implementation

[0020] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0021] like Figure 1 and Figure 2 As shown, this invention provides an embodiment of an online detection method for the impurity removal efficiency of a fluid iron separator, comprising: High-frequency electromagnetic induction sensors are deployed in the inlet and outlet pipes of the fluid separator to measure the instantaneous content of magnetic impurities in the fluid in real time, thereby obtaining the inlet impurity content sequence and the outlet impurity content sequence. A miniature fluxgate sensor array is deployed on the surface of the magnetic medium inside the fluid separator. The change in magnetic flux on the surface of the magnetic medium is monitored in real time by the miniature fluxgate sensor array, and the magnetic flux decay curve is obtained. The edge computing gateway receives the inlet impurity content sequence, the outlet impurity content sequence, and the magnetic flux attenuation curve. The edge computing gateway calculates the apparent impurity removal efficiency based on the inlet impurity content sequence and the outlet impurity content sequence, and calculates the magnetic medium adsorption saturation based on the magnetic flux decay curve. The edge computing gateway uses the magnetic medium adsorption saturation as the weight of adsorption effectiveness to perform weighted fusion of apparent impurity removal efficiency. When the magnetic medium adsorption saturation is lower than the preset saturation threshold, the confidence of apparent impurity removal efficiency is reduced and the apparent impurity removal efficiency is attenuated according to the ratio of the actual impurity removal efficiency to the apparent impurity removal efficiency multiplied by the magnetic medium adsorption saturation to the preset saturation threshold, so as to obtain the actual impurity removal efficiency.

[0022] In this embodiment, the fluid separator inlet pipe refers to the pipe section through which the fluid to be treated flows before entering the separator, and the fluid separator outlet pipe refers to the pipe section through which the fluid flows out after being treated by the separator. These two pipe sections are used to install detection sensors to obtain impurity content data of the fluid before entering the separator and after leaving the separator.

[0023] In this embodiment, the high-frequency electromagnetic induction sensor is a detection device based on the principle of electromagnetic induction. It emits a high-frequency electromagnetic field to induce eddy currents in magnetic impurities flowing through the sensor, thereby causing a change in the electromagnetic field. The sensor detects this change and converts it into an electrical signal output. The operating frequency of this sensor is typically above several kilohertz, enabling continuous and real-time detection of magnetic impurities in fluids.

[0024] In this embodiment, the real-time measurement of the instantaneous content of magnetic impurities in the fluid by a high-frequency electromagnetic induction sensor means that the sensor operates continuously and without interruption to detect the content of magnetic impurities in the fluid flowing through the sensor at every moment, and outputs the detection results in the form of electrical signals in real time, forming a detection data stream that changes continuously over time, thereby capturing the dynamic change process of impurity content.

[0025] In this embodiment, the instantaneous content of magnetic impurities refers to the amount of ferromagnetic impurities contained in a unit volume or unit mass of fluid at a specific moment. It is usually characterized by milligrams per liter or parts per million concentration. This value reflects the real-time concentration level of magnetic impurities in the fluid at that moment.

[0026] In this embodiment, the inlet impurity content sequence refers to a set of data arranged in chronological order, consisting of a series of instantaneous magnetic impurity content data measured by a high-frequency electromagnetic induction sensor at the inlet pipe of the fluid separator over a continuous time series. The outlet impurity content sequence refers to a set of data arranged in chronological order, consisting of a series of instantaneous magnetic impurity content data measured by a high-frequency electromagnetic induction sensor at the outlet pipe of the fluid separator over a continuous time series. These two sequences are used for subsequent calculations of the magnetic impurity removal effect of the separator.

[0027] In this embodiment, the magnetic medium surface inside the fluid iron separator refers to the outer surface of the magnetic element inside the iron separator used to adsorb magnetic impurities. These magnetic elements are usually made of high magnetic permeability material. When the fluid containing magnetic impurities flows through its surface, the impurities are adsorbed and accumulated on the surface under the action of the magnetic field to form an adsorption layer.

[0028] In this embodiment, the miniature fluxgate sensor array is a detection network composed of multiple miniature fluxgate sensors arranged in a certain spatial layout. Each miniature fluxgate sensor operates based on the fluxgate principle and can accurately measure the magnetic induction intensity at its location. This sensor array is embedded or attached to the surface of a magnetic medium to acquire information on the magnetic flux distribution at different locations on the surface of the magnetic medium.

[0029] In this embodiment, a miniature fluxgate sensor array monitors the changes in magnetic flux on the surface of the magnetic medium in real time. The obtained magnetic flux decay curve refers to the magnetic flux data collected by the miniature fluxgate sensor array at various detection points on the surface of the magnetic medium through continuous operation. The edge computing gateway then aggregates and processes this data to generate one or more curves reflecting the change in magnetic flux on the surface of the magnetic medium over time. As magnetic impurities are continuously adsorbed and accumulated on the surface of the magnetic medium, the magnetic flux on the surface of the magnetic medium will show a gradual decreasing trend.

[0030] In this embodiment, the magnetic flux decay curve is a curve plotted with time as the horizontal axis and magnetic flux value as the vertical axis. It is used to describe the change law of the magnetic flux on the surface of the magnetic medium gradually decreasing as the impurity adsorption process occurs. The shape and rate of change of the curve reflect the adsorption state and saturation process of the magnetic medium.

[0031] In this embodiment, the edge computing gateway is an intelligent device deployed in an industrial field, possessing data acquisition, storage, computing, and communication functions. This device receives real-time detection data from various sensors, performs data processing and analysis locally, and outputs or uploads the calculation results to the upper-level management system, offering advantages such as low latency and high reliability.

[0032] In this embodiment, calculating the apparent impurity removal efficiency based on the import impurity content sequence and the export impurity content sequence means that the edge computing gateway compares the import impurity content value and the export impurity content value at the same time point or within the same time window, and calculates the impurity removal efficiency value at that time point or time window by dividing the difference between the import impurity content value and the export impurity content value by the import impurity content value and then multiplying it by 100%. This efficiency value reflects the difference in quantity between import impurities and export impurities.

[0033] In this embodiment, the apparent impurity removal efficiency is a value directly calculated from the inlet impurity content sequence and the outlet impurity content sequence. This value only reflects the difference in the quantity of impurities at the inlet and outlet, but it cannot distinguish whether the reduction in outlet impurity content is due to the effective retention of impurities by the magnetic medium or because the impurities only pass through the magnetic medium area but are not detected.

[0034] In this embodiment, the magnetic medium adsorption saturation is a quantitative index used to characterize the degree of closeness between the current adsorption state of the magnetic medium and the fully saturated state. Its value range is usually between 0 and 1. The closer the value is to 1, the closer the adsorption capacity of the magnetic medium is to the saturated state. The closer the value is to 0, the more sufficient the adsorption capacity of the magnetic medium is.

[0035] In this embodiment, the adsorption effectiveness weight refers to the coefficient used to adjust the confidence level of the apparent impurity removal efficiency when calculating the impurity removal efficiency. This coefficient is determined based on the magnetic medium adsorption saturation and reflects the likelihood that impurities are effectively retained under the current state. The lower the adsorption saturation, the smaller the adsorption effectiveness weight, and the lower the confidence level of the apparent impurity removal efficiency.

[0036] In this embodiment, the edge computing gateway uses the magnetic medium adsorption saturation as a weight for the apparent impurity removal efficiency to perform weighted fusion. This means that the edge computing gateway compares the magnetic medium adsorption saturation with a preset saturation threshold. When the magnetic medium adsorption saturation is lower than the preset saturation threshold, the edge computing gateway reduces the confidence level of the apparent impurity removal efficiency and performs attenuation correction to obtain the true impurity removal efficiency. The basic principle of the correction is: the lower the magnetic medium adsorption saturation, the more impurities the magnetic medium has adsorbed and the weaker the remaining adsorption capacity. At this time, the possibility of impurities penetrating the iron separator is greater, and the more artificially high the apparent impurity removal efficiency is caused by penetrating impurities. Therefore, it is necessary to attenuate according to the ratio between the adsorption saturation and the preset saturation threshold.

[0037] The specific correction method is as follows: Divide the magnetic medium adsorption saturation by a preset saturation threshold to obtain the attenuation coefficient. When the magnetic medium adsorption saturation equals the preset saturation threshold, the attenuation coefficient is 1, and the apparent impurity removal efficiency does not decrease; when the magnetic medium adsorption saturation approaches 0, the attenuation coefficient approaches 0, and the apparent impurity removal efficiency is significantly attenuated. Multiplying the apparent impurity removal efficiency by this attenuation coefficient yields the true impurity removal efficiency.

[0038] The preset saturation threshold is determined as follows: Through continuous operation tests on fluid separators of the same specifications, the change in actual impurity removal efficiency is monitored as the magnetic medium adsorption saturation decreases. When the actual impurity removal efficiency begins to decrease significantly (e.g., the decrease exceeds 5%), the corresponding magnetic medium adsorption saturation value is recorded. This value is then multiplied by a safety factor (0.9 in this example) and used as the preset saturation threshold. For example, if the adsorption saturation value corresponding to the significant decrease in actual impurity removal efficiency is 70%, then the preset saturation threshold is 63%. This threshold can be adjusted according to the characteristics and process requirements of different types of separators.

[0039] In this embodiment, the preset saturation threshold is a pre-defined critical value for magnetic medium adsorption saturation, used to determine whether the magnetic medium has entered a state of significantly reduced adsorption capacity. This threshold is determined based on factors such as the design parameters of the iron separator, the characteristics of the magnetic medium material, and process requirements. When the actual monitored magnetic medium adsorption saturation is lower than this threshold, it indicates that the magnetic medium is close to saturation and the apparent impurity removal efficiency needs to be corrected.

[0040] In this embodiment, the true impurity removal efficiency is the value obtained after correction for the magnetic medium adsorption saturation. This value eliminates the deviation between apparent removal and actual retention caused by the decrease in the magnetic medium adsorption capacity, and can more accurately reflect the actual capture and retention effect of the fluid iron separator on magnetic impurities in the current state.

[0041] Furthermore, the step of calculating the magnetic medium adsorption saturation based on the magnetic flux decay curve can be implemented using one of the following two methods, the specific choice of which can be determined based on the on-site working conditions and the required detection accuracy: The first implementation method is suitable for application scenarios that require real-time capture of dynamic changes in the adsorption process and are sensitive to inflection points, and includes the following steps: The edge computing gateway performs a first-order difference operation on the magnetic flux decay curve to obtain the magnetic flux decay rate curve. The edge computing gateway performs a second-order difference operation on the magnetic flux decay rate curve to obtain the magnetic flux decay acceleration curve. The edge computing gateway searches for the moment when the magnetic flux decay rate approaches a preset minimum rate threshold in the magnetic flux decay acceleration curve. This moment is the turning point when the magnetic medium surface adsorption layer enters the critical state of collapse from a stable adsorption state. The preset minimum rate threshold is set to 0.01 millitalas per second. The edge computing gateway uses the magnetic flux value at the turning point as the initial saturation magnetic flux, and the ratio of the current magnetic flux to the initial saturation magnetic flux as the magnetic medium adsorption saturation. This ratio ranges from 0 to 1, and the closer the ratio is to 1, the closer the magnetic medium is to saturation.

[0042] The second implementation method is suitable for application scenarios that require trend assessment of the adsorption process and attention to the overall decay law, and includes the following steps: The edge computing gateway acquires the magnetic flux values ​​of multiple sampling points within a preset historical time period for the magnetic flux decay curve; The edge computing gateway performs exponential function fitting on the magnetic flux values ​​of multiple sampling points to obtain the magnetic flux decay fitting function; The edge computing gateway extracts the decay time constant of the magnetic flux decay fitting function. The ratio of the decay time constant to a preset reference time constant is used as the magnetic medium adsorption saturation. This ratio ranges from 0 to 1; the closer the ratio is to 1, the closer the magnetic medium's adsorption capacity is to a brand-new state. The preset reference time constant is obtained by calibrating a brand-new fluid separator of the same specification under standard operating conditions through adsorption experiments.

[0043] It should be noted that both implementation methods described above can effectively calculate the magnetic medium adsorption saturation, but the physical meanings of saturation defined by the two methods differ: the first method uses the inflection point as a reference, reflecting how close the current magnetic flux is to the saturation critical point; the second method uses a completely new state as a reference, reflecting how much the current decay rate has slowed down compared to the initial state. In practical applications, one method should be selected and used consistently throughout the process to ensure the consistency of the saturation calculation results. The system defaults to the first implementation method; when trend analysis based on historical data is required, the second implementation method can be switched to.

[0044] In this embodiment, performing a first-order difference operation on the magnetic flux decay curve to obtain the magnetic flux decay rate curve involves the edge computing gateway calculating the difference between the magnetic flux values ​​at two adjacent time points on the magnetic flux decay curve. The magnetic flux value at the later time point is subtracted from the magnetic flux value at the previous time point to obtain the change in magnetic flux within that time interval. This change is then divided by the time interval to obtain the average rate of change within that time interval. This same calculation is performed sequentially for all adjacent time points, and the results are arranged chronologically to obtain a magnetic flux decay rate curve with time on the horizontal axis and the rate of change of magnetic flux on the vertical axis. This curve reflects the rate at which magnetic flux decreases over time; a larger rate value indicates a faster decrease in magnetic flux.

[0045] In this embodiment, performing a second-order difference operation on the magnetic flux decay rate curve to obtain the magnetic flux decay acceleration curve involves the edge computing gateway calculating the difference between the rate values ​​at two adjacent time points on the magnetic flux decay rate curve. The rate value at the later time point is subtracted from the rate value at the previous time point to obtain the rate change within that time interval. This change is then divided by the time interval to obtain the average rate of change within that time interval. This same calculation is performed sequentially for all adjacent time points, and the results are arranged in chronological order to obtain a magnetic flux decay acceleration curve with time as the horizontal axis and magnetic flux acceleration as the vertical axis. This curve reflects the changing trend of the magnetic flux decay rate itself; a positive acceleration value indicates an accelerating decay rate, and a negative acceleration value indicates a decelerating decay rate.

[0046] In this embodiment, the preset minimum rate threshold is a pre-defined critical value for the magnetic flux decay rate, used to determine whether the magnetic flux decrease process has entered an extremely slow stage. This threshold is determined based on the material properties, structural dimensions, and a large amount of experimental data of the magnetic medium. In this embodiment, it is set to 0.01 millitalas per second. When the magnetic flux decay rate decreases to near this value, it indicates that the decrease in magnetic flux has become extremely slow, and the adsorption process is about to undergo a qualitative change.

[0047] In this embodiment, searching for the moment when the magnetic flux decay rate approaches a preset minimum rate threshold in the magnetic flux decay acceleration curve as the turning point when the magnetic medium surface adsorption layer transitions from a stable adsorption state to a critical collapse state means that the edge computing gateway first finds a time point on the magnetic flux decay rate curve where the rate value is close to 0.01 millitalas per second, and then verifies it on the magnetic flux decay acceleration curve near that time point. When it is confirmed that the acceleration value near this moment shows a regular change, this moment is determined as the critical time point when the state of the magnetic medium surface adsorption layer changes. This moment marks the transition of the magnetic medium adsorption process from one stable mode to another unstable mode.

[0048] In this embodiment, the stable adsorption state refers to the stage when the adsorption layer on the surface of the magnetic medium is in normal operation. At this time, magnetic impurities are continuously adsorbed onto the surface of the magnetic medium, the structure of the adsorption layer remains intact and stable, the magnetic flux shows a relatively gentle downward trend as impurities accumulate, the attenuation rate is relatively uniform, and the system can maintain normal impurity removal function.

[0049] In this embodiment, the collapse critical state refers to the stage where the adsorption layer on the surface of the magnetic medium is about to undergo structural changes. At this time, the adsorption layer has accumulated a large amount of impurities, its ability to continue adsorption is significantly weakened, the rate of decrease in magnetic flux is significantly slowed down, and the adsorption layer is on the verge of transitioning from stable accumulation to large-scale shedding. Once this state is entered, the adsorption layer may collapse locally at any time, causing impurities to be released again.

[0050] In this embodiment, using the magnetic flux value corresponding to the inflection point as the initial saturation magnetic flux means that after determining the inflection point, the edge computing gateway extracts the magnetic flux value corresponding to that moment from the original magnetic flux decay curve and records this value as a benchmark reference value for subsequent calculations. This value represents the magnetic flux level of the magnetic medium before it enters an unstable state and is used to assess the degree of decrease in the current magnetic flux relative to this benchmark value.

[0051] This embodiment employs differential analysis, performing first- and second-order difference operations on the magnetic flux decay curve to identify the inflection point where the magnetic flux decay rate approaches a preset minimum rate threshold. The magnetic flux value corresponding to this point is taken as the initial saturation magnetic flux, and the ratio of the current magnetic flux to the initial saturation magnetic flux is calculated to obtain the adsorption saturation. This method focuses on capturing the dynamic changes in the adsorption process, determining the saturation benchmark through inflection point identification. It is suitable for applications requiring real-time capture of abrupt changes in the adsorption state, enabling timely response to changes in the magnetic medium entering a critical state of collapse.

[0052] In this embodiment, obtaining the magnetic flux values ​​of multiple sampling points within a preset historical time period from the magnetic flux decay curve refers to the edge computing gateway taking the current time as the endpoint and tracing back the preset historical time period length to extract the magnetic flux values ​​corresponding to multiple time points within that time period from the stored magnetic flux decay curve. These sampling points are selected at fixed time intervals, such as one point every one minute or ten minutes, forming a set of magnetic flux data sequences arranged in chronological order for subsequent mathematical analysis.

[0053] In this embodiment, the magnetic flux values ​​at multiple sampling points are fitted with an exponential function to obtain the magnetic flux decay fitting function. This means that the edge computing gateway uses multiple time points collected within a preset historical period and their corresponding magnetic flux values ​​as input data, employs an exponential function as the fitting model, and uses mathematical optimization algorithms such as the least squares method to calculate an exponential curve that best describes the distribution pattern of these data points. The mathematical form of this exponential curve is that the magnetic flux value equals the initial magnetic flux multiplied by the negative decay coefficient of the natural constant multiplied by the power of time. After fitting, a specific function expression is obtained, which can describe the overall law of magnetic flux decay over time.

[0054] In this embodiment, extracting the decay time constant of the magnetic flux decay fitting function refers to taking the parameters related to the decay coefficient from the obtained exponential function expression and converting them to obtain the characteristic value of the decay time constant. The decay time constant is the core parameter describing the speed of the exponential decay process. The larger the value, the slower the magnetic flux decays and the longer the magnetic medium's adsorption capacity is maintained; the smaller the value, the faster the magnetic flux decays and the more rapidly the magnetic medium's adsorption capacity decreases.

[0055] In this embodiment, the preset reference time constant is a pre-set reference value used for comparison with the currently extracted decay time constant. This reference value represents the decay time constant that a brand-new magnetic medium should have under standard operating conditions, reflecting the adsorption performance characteristics of the magnetic medium in its ideal initial state. Different types of magnetic media and different specifications of iron separators have different reference time constants.

[0056] In this embodiment, the preset reference time constant is obtained by calibrating a brand-new fluid separator of the same specification under standard operating conditions. This means that before actual application, a brand-new fluid separator of the exact same model and specifications as the one used on-site is selected, placed under standard operating conditions, and a standard test fluid containing known concentrations of magnetic impurities is introduced. The magnetic flux decay process is continuously monitored until saturation. By collecting the magnetic flux decay data throughout the process and performing exponential function fitting, the decay time constant of the brand-new device under ideal conditions is extracted. This value is recorded as the preset reference time constant for comparison during subsequent actual operation.

[0057] This embodiment employs an exponential fitting method, fitting an exponential function to the magnetic flux decay curve at multiple sampling points within a preset historical time period. The decay time constant is extracted, and its ratio to a preset baseline time constant is used as the adsorption saturation. This method focuses on describing the overall trend of magnetic flux decay, quantifying the adsorption state through time constant comparison. It is suitable for applications that assess the trend of the adsorption process and can predict the evolution of the adsorption state based on historical data.

[0058] Furthermore, it also includes steps for correcting the uniformity of the entire flow field: As an optional optimization step, after calculating the true impurity removal efficiency based on the magnetic medium adsorption saturation, a full flow field uniformity correction can be further introduced to eliminate detection deviations caused by uneven fluid distribution. This correction step is independent of the aforementioned adsorption saturation correction, and the two are performed sequentially. That is, after obtaining the true impurity removal efficiency through adsorption saturation correction, the flow field uniformity coefficient is used as a secondary correction factor.

[0059] The specific steps for correcting the uniformity of the entire flow field are as follows: Distributed magnetic sensors are deployed at the inlet section, the middle section, and the outlet section of the internal flow channel of the fluid separator. The magnetic impurity concentration data at different spatial locations are collected synchronously through the distributed magnetic sensors. The edge computing gateway receives magnetic impurity concentration data collected by the distributed magnetic sensor and calculates the axial concentration gradient based on the magnetic impurity concentration data of the inlet section, the middle section, and the outlet section. The edge computing gateway uses the axial concentration gradient as the flow field uniformity coefficient. When the flow field uniformity coefficient exceeds the preset uniformity threshold, the actual impurity removal efficiency obtained after adsorption saturation correction is corrected a second time to obtain the final actual impurity removal efficiency used for output.

[0060] It should be noted that the overall flow field uniformity correction is an optional step. If the on-site operating conditions are stable and the flow field distribution is uniform, this correction can be omitted. When this correction is enabled, the final output of the true impurity removal efficiency is the result after dual processing of adsorption saturation correction and flow field uniformity correction.

[0061] In this embodiment, the inlet cross-section of the internal flow channel of the fluid separator refers to the plane position perpendicular to the flow direction that the fluid passes through when it first enters the separator; the middle cross-section refers to the plane position perpendicular to the flow direction that the fluid passes through when it flows through the middle region of the separator; and the outlet cross-section refers to the plane position perpendicular to the flow direction that the fluid passes through when it is about to leave the separator. These three cross-sections are distributed sequentially along the fluid flow direction to collect magnetic impurity distribution data at different spatial locations, reflecting the concentration changes of impurities inside the separator.

[0062] In this embodiment, the distributed magnetic sensor is a sensor network composed of multiple magnetic detection units arranged in a certain spatial layout. Each detection unit operates based on the principle of electromagnetic induction and can measure the concentration of magnetic impurities at its location. These detection units are deployed at different spatial points on the inlet, middle, and outlet sections of the flow channel inside the iron separator to obtain information on the concentration of magnetic impurities at different locations at the same time.

[0063] In this embodiment, the synchronous acquisition of magnetic impurity concentration data at different spatial locations via distributed magnetic sensors means that all magnetic detection units deployed at the inlet, middle, and outlet sections start measuring simultaneously, acquiring the magnetic impurity concentration value at their respective locations, and transmitting this data to the edge computing gateway. This synchronous acquisition method can accurately reflect the impurity distribution at different locations inside the iron separator at a given moment, avoiding the data incomparability problem caused by time differences.

[0064] In this embodiment, the magnetic impurity concentration data refers to the numerical value obtained by the distributed magnetic sensor, which reflects the amount of magnetic impurities contained in a unit volume or unit mass of fluid at the location. It is usually characterized by milligrams per liter or parts per million. This value is used to analyze the distribution pattern of impurities at different locations in the flow channel inside the iron separator.

[0065] In this embodiment, calculating the axial concentration gradient based on the magnetic impurity concentration data of the inlet section, the middle section, and the outlet section involves the edge computing gateway subtracting the average concentration value of the outlet section from the average concentration value of the inlet section to obtain the overall concentration decrease along the flow direction. This is then combined with the distance between the inlet and outlet sections to calculate the concentration change rate per unit length. Simultaneously, subtracting the average concentration value of the middle section from the average concentration value of the inlet section, and subtracting the average concentration value of the outlet section from the average concentration value of the middle section, yields the front-end and rear-end concentration gradients, respectively, used to analyze the distribution of impurities in different sections within the iron separator during the impurity removal process.

[0066] In this embodiment, the flow field uniformity coefficient is an index used to quantitatively characterize the uniformity of fluid distribution inside the iron separator, and its value is calculated based on the axial concentration gradient. The smaller the axial concentration gradient, the more gradual the change in impurity concentration along the flow direction, and the more uniform the fluid distribution; the larger the axial concentration gradient, the more drastic the change in impurity concentration along the flow direction, the more uneven the fluid distribution, and the possible existence of local penetration channels or short-circuit flows.

[0067] In this embodiment, the preset gradient threshold is a pre-set critical value for axial concentration gradient, used to determine whether the internal flow field distribution of the iron separator is within an acceptable uniform range. This threshold is determined based on factors such as the structural design parameters of the iron separator, fluid characteristics, and process requirements. When the actual monitored axial concentration gradient exceeds this threshold, it indicates that the flow field distribution is significantly uneven, and a secondary correction to the actual impurity removal efficiency is required.

[0068] Furthermore, the steps of incorporating the axial concentration gradient as a flow field uniformity coefficient into the calculation of the actual impurity removal efficiency include: The edge computing gateway acquires magnetic impurity concentration data for fine particle size range, medium particle size range, and coarse particle size range collected by distributed magnetic sensors. The edge computing gateway calculates the axial concentration gradients for the fine particle size range, the medium particle size range, and the coarse particle size range, respectively. The edge computing gateway performs a weighted summation of the axial concentration gradients of the fine particle size range, the medium particle size range, and the coarse particle size range according to preset weights to obtain the particle size weighted uniformity coefficient. When the particle size weighted uniformity coefficient exceeds the preset uniformity threshold, the edge computing gateway performs a second correction on the actual impurity removal efficiency. Finally, the actual impurity removal efficiency is equal to the actual impurity removal efficiency multiplied by the ratio of the preset uniformity threshold to the particle size weighted uniformity coefficient.

[0069] In this embodiment, the fine particle size range refers to the range of smaller magnetic impurities defined according to fluid characteristics and iron separator design requirements. Magnetic impurity particles in this range have smaller diameters and are more likely to move with the fluid in the flow field, easily being carried by the fluid to penetrate the magnetic medium region. Their impact on impurity removal efficiency differs from that of larger particles.

[0070] In this embodiment, the magnetic impurity concentration data in the fine particle size range refers to the concentration value obtained by the distributed magnetic sensor by adjusting the electromagnetic excitation frequency to specifically detect magnetic impurities in the fine particle size range. This data reflects the distribution of fine particle magnetic impurities at a specific spatial location.

[0071] In this embodiment, the medium particle size range refers to the range of intermediate-sized magnetic impurities defined according to fluid characteristics and iron separator design requirements. The diameter of magnetic impurity particles in this range is between that of fine and coarse particles, and their movement behavior in the flow field and adsorption characteristics on magnetic media have their own characteristics.

[0072] In this embodiment, the magnetic impurity concentration data in the medium particle size range refers to the concentration value obtained by the distributed magnetic sensor by adjusting the electromagnetic excitation frequency to specifically detect magnetic impurities in the medium particle size range. This data reflects the distribution of medium particle magnetic impurities at a specific spatial location.

[0073] In this embodiment, the coarse particle size range refers to the range of larger magnetic impurities defined according to fluid characteristics and iron separator design requirements. Magnetic impurity particles in this range have larger diameters and greater inertia in the flow field, making them easier to be captured by the magnetic medium, but they may also deposit unevenly on the surface of the magnetic medium.

[0074] In this embodiment, the magnetic impurity concentration data in the coarse particle size range refers to the concentration value obtained by the distributed magnetic sensor by adjusting the electromagnetic excitation frequency to specifically detect magnetic impurities in the coarse particle size range. This data reflects the distribution of coarse magnetic impurities at a specific spatial location.

[0075] In this embodiment, calculating the axial concentration gradients for the fine particle size range, the medium particle size range, and the coarse particle size range separately means that the edge computing gateway independently calculates the axial concentration gradient for each particle size range. For the fine particle size range, the axial concentration gradient is calculated using concentration data from the inlet, middle, and outlet sections of this range; for the medium particle size range, the axial concentration gradient is calculated using concentration data from these three sections; and the coarse particle size range is also calculated independently. This yields three gradient values ​​that reflect the uniformity of impurity distribution along the flow direction for different particle sizes.

[0076] In this embodiment, the edge computing gateway performs a weighted summation of the axial concentration gradients of the fine particle size range, the medium particle size range, and the coarse particle size range according to preset weights to obtain a particle size weighted uniformity coefficient. The preset weights are determined based on the degree of influence and removal difficulty of impurities in each particle size range on downstream equipment. For example, the weight of the fine particle size range is 30%, the weight of the medium particle size range is 40%, and the weight of the coarse particle size range is 30%, with the sum of the three weights being 100%. The criteria for dividing the fine, medium, and coarse particle size ranges are determined based on the design parameters and process requirements of the fluid iron separator. For example, the fine particle size range is 0 to 20 micrometers, the medium particle size range is 20 to 50 micrometers, and the coarse particle size range is above 50 micrometers.

[0077] In this embodiment, the axial concentration gradients of the fine particle size range, the medium particle size range, and the coarse particle size range are weighted and summed according to preset weights to obtain the particle size weighted uniformity coefficient. This is achieved by the edge computing gateway multiplying the axial concentration gradient of each particle size range by its corresponding preset weight, and then summing the three products to obtain a comprehensive value. This value considers both the uniformity of the distribution of impurities of different sizes in the flow field and the different influence weights of each particle size range on the overall performance.

[0078] In this embodiment, the particle size weighted uniformity coefficient is a comprehensive quantitative index used to characterize the overall uniformity of the flow field after considering particle size differences. This coefficient integrates the axial distribution characteristics of impurities in three particle size ranges: fine, medium, and coarse, and weights them according to the importance of each particle size range, thus more comprehensively reflecting the distribution of impurities of different sizes inside the iron separator.

[0079] In this embodiment, the preset uniformity threshold is a pre-set critical value of the particle size weighted uniformity coefficient, used to determine whether the uniformity of the flow field after considering particle size differences is within an acceptable range. This threshold is determined comprehensively based on factors such as the design parameters of the iron separator, process requirements, and control targets for impurities of different particle sizes. For example, this threshold is set to 0.8, calibrated by collecting 20 sets of data under uniform flow field distribution (axial concentration gradient less than or equal to 0.1 mg / L / m) and 20 sets of data under non-uniform flow field distribution (axial concentration gradient greater than or equal to 0.5 mg / L / m). The particle size weighted uniformity coefficient for each set of data is calculated, and the statistical analysis shows that the dividing point of the coefficient distribution between the two sets of data is between 0.75 and 0.85. The median value of 0.8 is taken as the preset uniformity threshold. This value indicates that when the particle size weighted uniformity coefficient exceeds 0.8, the flow field distribution is significantly non-uniform, and the impact of flow field non-uniformity on efficiency detection error exceeds 10%, requiring secondary correction of the actual impurity removal efficiency.

[0080] In this embodiment, the actual impurity removal efficiency is corrected a second time. The final actual impurity removal efficiency is equal to the ratio of the actual impurity removal efficiency multiplied by the preset uniformity threshold to the particle size weighted uniformity coefficient. This means that when the particle size weighted uniformity coefficient exceeds the preset uniformity threshold, the edge computing gateway initiates a correction program, multiplying the current actual impurity removal efficiency by the ratio of the preset uniformity threshold to the particle size weighted uniformity coefficient to obtain the corrected new efficiency value. Since a larger particle size weighted uniformity coefficient indicates a more non-uniform flow field distribution, this ratio is less than one. The corrected efficiency value is lower than the original value, thus deducting the overestimation of efficiency caused by non-uniform flow field, making the final detection result more accurate and reliable.

[0081] Furthermore, a dynamic compensation step is included before the step of calculating the apparent impurity removal efficiency based on the import impurity content sequence and the export impurity content sequence: A flow velocity sensor and a viscosity sensor are deployed in the inlet pipe of the fluid separator. The flow velocity sensor collects the fluid flow velocity in real time, and the viscosity sensor collects the fluid viscosity in real time. The edge computing gateway receives fluid velocity and fluid viscosity, and determines the dynamic compensation coefficient according to the ratio of fluid velocity to reference velocity and the ratio of fluid viscosity to reference viscosity in a linear function relationship. The dynamic compensation coefficient is equal to 1 plus the difference between the velocity compensation coefficient and the ratio of fluid velocity to reference velocity minus 1, plus the difference between the viscosity compensation coefficient and the ratio of fluid viscosity to reference viscosity minus 1. The edge computing gateway corrects the instantaneous magnetic impurity content in the inlet impurity content sequence and the instantaneous magnetic impurity content in the outlet impurity content sequence according to the dynamic compensation coefficient. The corrected instantaneous magnetic impurity content is equal to the original instantaneous magnetic impurity content multiplied by the dynamic compensation coefficient. The edge computing gateway uses the corrected inlet impurity content sequence and the corrected outlet impurity content sequence as input data to calculate the apparent impurity removal efficiency.

[0082] In this embodiment, the flow velocity sensor refers to a flow measurement device deployed on the inlet pipe of the fluid separator, used to detect the flow velocity of the fluid in the pipe in real time. It is usually an electromagnetic flow meter, ultrasonic flow meter or turbine flow meter, etc., and can continuously output the instantaneous value of the fluid flow velocity. The viscosity sensor refers to a fluid viscosity measurement device deployed on the inlet pipe of the fluid separator, used to detect the viscosity of the fluid in real time. It is usually a vibratory viscometer or capillary viscometer, etc., and can continuously output the instantaneous value of the fluid viscosity.

[0083] In this embodiment, the fluid velocity is collected in real time by a flow velocity sensor, and the fluid viscosity is collected in real time by a viscosity sensor. This means that the two sensors work continuously and without interruption to measure the fluid velocity and viscosity flowing through the inlet pipe at every moment, and output the measurement results in real time in the form of electrical signals, forming a flow velocity data stream and a viscosity data stream that change continuously over time, which are then used by the edge computing gateway for subsequent dynamic compensation calculations.

[0084] In this embodiment, the reference flow rate refers to a pre-set reference flow rate value used for comparison with the real-time collected fluid flow rate. This reference value is usually determined based on the design conditions of the iron separator, sensor calibration conditions, or process specifications, reflecting the flow rate level that the fluid should have under ideal or standard conditions.

[0085] In this embodiment, the reference viscosity refers to a pre-set reference viscosity value used for comparison with the fluid viscosity collected in real time. This reference value is usually determined based on the design conditions of the iron separator, sensor calibration conditions, or process specifications, reflecting the viscosity level that the fluid should have under ideal or standard conditions.

[0086] In this embodiment, the linear function relationship means that the calculation of the dynamic compensation coefficient adopts the mathematical form of a linear function. That is, the compensation coefficient and each input variable maintain a proportional increase or decrease relationship, without involving higher-order terms or complex nonlinear operations, making the calculation process of the compensation coefficient simple and intuitive, and convenient for engineering implementation and on-site debugging.

[0087] In this embodiment, the dynamic compensation coefficient equals 1 plus the difference between the velocity compensation coefficient multiplied by the ratio of the fluid velocity to the reference velocity minus 1, plus the difference between the viscosity compensation coefficient multiplied by the ratio of the fluid viscosity to the reference viscosity minus 1. This means that the edge computing gateway calculates the dynamic compensation coefficient according to the following steps: First, calculate the ratio of the fluid velocity to the reference velocity, and subtract one from this ratio to obtain the velocity deviation; multiply the velocity compensation coefficient by this velocity deviation to obtain the velocity contribution value; then calculate the ratio of the fluid viscosity to the reference viscosity, and subtract one from this ratio to obtain the viscosity deviation; multiply the viscosity compensation coefficient by this viscosity deviation to obtain the viscosity contribution value; finally, add one to the velocity contribution value and the viscosity contribution value to obtain the dynamic compensation coefficient. This calculation formula ensures that when the velocity or viscosity deviates from the reference value, the compensation coefficient increases accordingly, and the greater the deviation, the greater the compensation.

[0088] In this embodiment, the flow velocity compensation coefficient refers to a pre-set weight value used to adjust the degree of influence of flow velocity on the dynamic compensation coefficient. The magnitude of this coefficient reflects the intensity of the influence of flow velocity changes on the accuracy of the detection results. The flow velocity compensation coefficient is determined based on sensor characteristics, fluid properties, and experimental data calibration, and its value range is typically between zero and one.

[0089] In this embodiment, the viscosity compensation coefficient refers to a pre-set weight value used to adjust the degree of influence of viscosity factors on the dynamic compensation coefficient. The magnitude of this coefficient reflects the intensity of the influence of viscosity changes on the accuracy of the detection results. The viscosity compensation coefficient is determined based on sensor characteristics, fluid properties, and experimental data calibration, and its value range is typically between zero and one.

[0090] Furthermore, the steps of obtaining the import impurity content sequence and the export impurity content sequence include a sub-step of multi-frequency electromagnetic response identification: The high-frequency electromagnetic induction sensor alternately emits electromagnetic waves at a first detection frequency and a second detection frequency, wherein the first detection frequency is lower than the second detection frequency. The edge computing gateway receives a first response signal corresponding to a first detection frequency and a second response signal corresponding to a second detection frequency. Edge computing gateways differentiate between strongly magnetic ferromagnetic impurities and weakly magnetic impurities based on the differences in their response characteristics to electromagnetic waves of different frequencies. Low-frequency electromagnetic waves have strong penetrating power and can simultaneously excite both strongly magnetic ferromagnetic and weakly magnetic impurities to generate response signals; high-frequency electromagnetic waves have weak penetrating power, and their signals are mainly affected by the surface characteristics of the impurities. Based on this principle, by comparing the response signals at the two frequencies, the contribution of weakly magnetic impurities can be separated. Specifically, when the first response signal corresponding to the first detection frequency is greater than or equal to the second response signal corresponding to the second detection frequency, the difference between the first response signal and the second response signal is taken as the contribution value of the weak magnetic impurity, and the remaining part after subtracting the contribution value of the weak magnetic impurity from the second response signal is taken as the content value of the strong magnetic ferromagnetic impurity. When the first response signal is less than the second response signal, it indicates that the contribution of weakly magnetic impurities is insignificant or that measurement interference exists. In this case, the contribution value of weakly magnetic impurities is set to zero, and the second response signal is directly used as the content value of strongly magnetic ferromagnetic impurities. This processing method avoids negative values ​​due to abnormal data and ensures the physical rationality of the detection results. The edge computing gateway regenerates the inlet and outlet impurity content sequences based on the content values ​​of strongly magnetic and ferromagnetic impurities, so that each content value in the inlet and outlet impurity content sequences corresponds to a strongly magnetic and ferromagnetic impurity that can be effectively removed by the fluid iron separator.

[0091] In this embodiment, the first detection frequency refers to the lower frequency value used when the high-frequency electromagnetic induction sensor emits electromagnetic waves. This lower frequency has a stronger electromagnetic wave penetration capability, which can simultaneously excite both strongly magnetic ferromagnetic impurities and weakly magnetic impurities in the fluid to generate response signals.

[0092] In this embodiment, the second detection frequency refers to the higher frequency value used when the high-frequency electromagnetic induction sensor emits electromagnetic waves. This frequency is higher, and the electromagnetic wave penetration ability is relatively weaker. According to the skin effect principle, its signal mainly reflects information close to the surface of the impurity. The response characteristics to strongly magnetic ferromagnetic impurities and weakly magnetic impurities are different from those of the first detection frequency.

[0093] In this embodiment, alternating the emission of electromagnetic waves at the first detection frequency and the second detection frequency means that during continuous operation, the high-frequency electromagnetic induction sensor alternately switches the emission frequency according to a set time interval. First, it emits electromagnetic waves at the first detection frequency and collects the response signal, and then switches to the second detection frequency to emit electromagnetic waves and collect the response signal. This process is repeated to obtain alternating detection data at the two frequencies.

[0094] In this embodiment, receiving the first response signal corresponding to the first detection frequency and the second response signal corresponding to the second detection frequency means that the edge computing gateway collects the electrical signal values ​​output by the sensor when it operates at the first detection frequency and the electrical signal values ​​output when it operates at the second detection frequency, respectively. These two values ​​reflect the comprehensive response intensity generated by all magnetic impurities in the fluid under different frequency excitation conditions.

[0095] In this embodiment, the skin effect principle refers to the phenomenon that when an alternating electromagnetic field propagates in a conductor, the current density concentrates near the conductor's surface. The higher the frequency, the more the current tends to distribute towards the surface, resulting in a shallower penetration depth. When applied to electromagnetic detection, electromagnetic waves of different frequencies probe the interior and surface of impurities to different depths, thereby allowing the differentiation of impurity types with different electromagnetic properties.

[0096] In this embodiment, the first response signal being greater than or equal to the second response signal means that the value of the response signal collected by the sensor when it operates at a lower frequency is not lower than the value of the response signal collected when it operates at a higher frequency. This situation conforms to normal physical laws and indicates that the detection data at the two frequencies can be used for subsequent impurity type differentiation calculations.

[0097] In this embodiment, the difference between the first response signal and the second response signal is used as the contribution value of the weak magnetic impurity. This means that the edge computing gateway subtracts the response signal value at a higher frequency from the response signal value at a lower frequency. The difference represents the part of the response that is mainly affected by the skin effect. This part of the response mainly comes from the weak magnetic impurity because the weak magnetic impurity is more sensitive to frequency changes.

[0098] In this embodiment, the remaining part after subtracting the contribution value of weak magnetic impurities from the second response signal is taken as the content value of strong magnetic ferromagnetic impurities. This means that the edge computing gateway subtracts the calculated contribution value of weak magnetic impurities from the response signal value at a higher frequency. The remaining value obtained represents the part of the response that is not significantly affected by the skin effect. This part of the response mainly comes from strong magnetic ferromagnetic impurities because strong magnetic ferromagnetic impurities are relatively insensitive to frequency changes.

[0099] In this embodiment, the first response signal being less than the second response signal means that the response signal value collected by the sensor when it operates at a lower frequency is lower than the response signal value collected when it operates at a higher frequency. This situation may be caused by flow field interference, signal noise, or special impurities, which does not conform to normal physical laws and requires adjustment of the data processing logic.

[0100] In this embodiment, the contribution value of weak magnetic impurities is set to zero, and the second response signal is directly used as the content value of strong magnetic ferromagnetic impurities. This means that when the first response signal is less than the second response signal, the edge computing gateway abandons the difference calculation method and directly uses the response signal value at a higher frequency as the content value of strong magnetic ferromagnetic impurities. This avoids the occurrence of negative values ​​or false increases due to data anomalies and ensures the physical rationality of the calculation results.

[0101] In this embodiment, the inlet and outlet impurity content sequences are regenerated based on the content values ​​of strongly magnetic ferromagnetic impurities. Each content value in both sequences corresponds to a strongly magnetic ferromagnetic impurity that can be effectively removed by the fluid separator. This means that after the edge computing gateway removes the contribution of weakly magnetic impurities from the original detection data, it replaces the original comprehensive impurity content value with a pure strongly magnetic ferromagnetic impurity content value, thus reconstructing the inlet and outlet impurity content time series. After this processing, the data used in subsequent calculations of apparent and true impurity removal efficiency only includes the strongly magnetic ferromagnetic impurities that can actually be captured by the separator, eliminating the interference of weakly magnetic impurities on the detection results. This makes the final impurity removal efficiency more accurately reflect the true working performance of the separator.

[0102] Furthermore, after obtaining the true impurity removal efficiency, a trend prediction step is also included: The edge computing gateway obtains the actual noise removal efficiency calculated at multiple time points within a historical period, forming a historical sequence of actual noise removal efficiency. The edge computing gateway performs time-series analysis on the historical sequence of the actual noise removal efficiency to extract the decay rate of the actual noise removal efficiency; The edge computing gateway predicts the trend of real noise removal efficiency in the future based on the decay rate of real noise removal efficiency. When the predicted real noise removal efficiency will drop below the target efficiency threshold within a preset time window, an efficiency decay warning signal is generated.

[0103] In this embodiment, the historical time period refers to a pre-defined time interval length that traces back from the current moment, used to limit the time range of the actual impurity removal efficiency data participating in trend analysis. The setting of this time period length needs to comprehensively consider the operating characteristics of the iron separator and the detection accuracy requirements, and is usually set to several hours to several days to ensure that the collected data can fully reflect the efficiency change pattern, while not introducing too many historical interference factors due to an excessively long time span.

[0104] In this embodiment, obtaining the actual noise removal efficiency calculated at multiple time points within a historical period and forming a historical sequence of actual noise removal efficiency means that the edge computing gateway, with the current time as the endpoint, traces back a preset historical period length and extracts the actual noise removal efficiency values ​​corresponding to multiple time points within that period from the stored detection data. These time points are selected at fixed time intervals, such as one point every ten minutes or one hour, and the selected efficiency values ​​are arranged in chronological order to form a data sequence for subsequent time-series analysis.

[0105] In this embodiment, time-series analysis is performed on the historical sequence of actual noise removal efficiency to extract the decay rate of actual noise removal efficiency. This refers to the edge computing gateway using time-series analysis methods to process the actual noise removal efficiency data sequence within a historical period and calculate the rate at which the efficiency value decreases over time. Specifically, a linear regression method can be used to fit a trend line of efficiency changing over time, and the slope of this trend line is the decay rate; or an exponential smoothing method can be used to calculate the weighted average rate of change of efficiency. The decay rate reflects the magnitude of the decrease in actual noise removal efficiency per unit time, expressed as a percentage per hour or per day.

[0106] In this embodiment, the future time period refers to a pre-defined time interval extending from the current moment into the future, used to define the time range for predicting the actual trend of changes in impurity removal efficiency. The length of this time period depends on the early warning response requirements and process control requirements, and is typically set to tens of minutes to several hours to allow sufficient response time before a significant drop in efficiency.

[0107] In this embodiment, predicting the future trend of actual noise removal efficiency based on the decay rate of the actual noise removal efficiency means that the edge computing gateway takes the current actual noise removal efficiency as the starting point, uses the extracted decay rate as the slope of change, and calculates the predicted efficiency value at each future time point using a linear extrapolation method. Specifically, the current efficiency value is successively subtracted from the decay rate multiplied by the time interval to obtain the efficiency estimate value at different future times, forming an efficiency prediction curve extending from the present to the future, reflecting the possible path of efficiency change in the future period.

[0108] In this embodiment, the preset time window is a pre-defined time length used to determine whether the efficiency decline will reach a critical level in the short term. This time window is typically set to a value adapted to the operation response time, such as thirty minutes or one hour. When the prediction results show that the efficiency will drop below the warning line within this time window, it means that the operator needs to take measures as soon as possible.

[0109] In this embodiment, the target efficiency threshold is a pre-set minimum allowable value for actual impurity removal efficiency, used to determine whether the iron separator is in a qualified working state. This threshold is determined based on downstream equipment protection requirements, product quality standards, and process specifications. When the actual impurity removal efficiency is lower than this threshold, it indicates that the iron separator can no longer meet normal production needs and requires timely cleaning or maintenance. For example, setting it to 85% means that when the actual impurity removal efficiency is lower than 85%, the iron separator can no longer meet normal production needs.

[0110] In this embodiment, the efficiency decay early warning signal refers to a warning message automatically generated by the edge computing gateway after completing trend prediction, when it finds that the prediction result indicates that the actual impurity removal efficiency will drop below the target efficiency threshold within a preset time window. This signal can be an alarm code in the Ethernet communication protocol, a prompt message on the display screen, a notification message sent to the operator's mobile terminal, or an early warning command connected to the factory control system, used to remind relevant personnel to pay attention to the status of the iron separator and arrange maintenance operations.

[0111] Furthermore, after obtaining the actual impurity removal efficiency, a step to indicate the cleaning effect is also included: The edge computing gateway obtains the historical cleaning records of the fluid iron separator. The historical cleaning records include the magnetic media adsorption saturation before each cleaning operation, the magnetic media adsorption saturation after each cleaning operation, and the actual impurity removal efficiency recovery value after each cleaning operation. The edge computing gateway establishes a cleaning effect prediction model based on historical cleaning records. The cleaning effect prediction model is used to characterize the relationship between the magnetic medium adsorption saturation and the actual cleanliness recovery value after cleaning. The edge computing gateway inputs the current magnetic medium adsorption saturation into the cleaning effect prediction model to obtain the predicted efficiency recovery value after the current cleaning operation, and generates cleaning timing suggestions and expected effect prompts based on the predicted efficiency recovery value, without correcting the current actual impurity removal efficiency.

[0112] In this embodiment, the historical cleaning record of the fluid magnetic separator includes the magnetic media adsorption saturation before each cleaning operation, the magnetic media adsorption saturation after each cleaning operation, and the actual impurity removal efficiency recovery value after each cleaning operation. This means that the edge computing gateway automatically records three key data points each time it performs a cleaning operation on the magnetic separator: the magnetic media adsorption saturation value monitored moment before the cleaning operation begins, the magnetic media adsorption saturation value monitored moment after the cleaning operation ends, and the difference in actual impurity removal efficiency before and after the cleaning operation. These data are stored chronologically to form a historical database of cleaning effects for subsequent analysis.

[0113] In this embodiment, the actual impurity removal efficiency recovery value refers to the increase in actual impurity removal efficiency relative to the pre-cleaning level after each cleaning operation. Specifically, it is calculated by subtracting the actual impurity removal efficiency before the cleaning operation from the actual impurity removal efficiency after the cleaning operation. This recovery value reflects the actual effect of the cleaning operation on restoring the performance of the iron separator; a larger recovery value indicates a more significant cleaning effect.

[0114] In this embodiment, a cleaning effect prediction model is established based on historical cleaning records. This model characterizes the relationship between magnetic media adsorption saturation and the actual recovery value of cleaning efficiency. The edge computing gateway performs statistical analysis on multiple sets of data accumulated in the historical cleaning records, using the magnetic media adsorption saturation before cleaning as the input variable and the actual recovery value of cleaning efficiency after cleaning as the output variable. Mathematical methods such as linear regression or exponential fitting are used to establish a functional relationship between the two. After the model is built, when any magnetic media adsorption saturation value is input, the model can output the corresponding expected efficiency recovery value. Specifically, the edge computing gateway collects at least 10 sets of historical cleaning record data. Each set of data includes the magnetic media adsorption saturation before cleaning as the independent variable x and the actual recovery value of cleaning efficiency after cleaning as the dependent variable y. A univariate linear regression method is used to fit the above data to obtain a regression equation. After fitting, the goodness-of-fit R-squared value of the model is calculated, which reflects the degree of agreement between the historical data and the regression model. When the R-squared value is greater than or equal to 0.8, the model is considered to have good predictive ability and can be put into use. When the R-squared value is less than 0.8, it indicates that the currently accumulated historical cleaning record data is insufficient to establish a reliable predictive model. The edge computing gateway continues to accumulate cleaning record data, and the model is only activated after the data volume is sufficient and the R-squared value reaches 0.8 or higher. For example, the threshold of 0.8 is determined based on conventional experience in industrial field data modeling, representing an acceptable level of agreement between the data and the model.

[0115] In this embodiment, inputting the current magnetic medium adsorption saturation into the cleaning effect prediction model to obtain the predicted efficiency recovery value after the current cleaning operation means that the edge computing gateway uses the currently monitored magnetic medium adsorption saturation value as input, substitutes it into the established cleaning effect prediction model, and calculates the expected actual improvement in impurity removal efficiency after performing a cleaning operation under the current adsorption state. This predicted value provides the operator with a quantitative expectation of the cleaning effect.

[0116] In this embodiment, the cleaning timing suggestion and expected effect prompt are generated based on the predicted efficiency recovery value. The fact that the current actual cleaning efficiency is not corrected means that the edge computing gateway compares the calculated predicted efficiency recovery value with a preset efficiency improvement threshold. When the predicted efficiency recovery value meets production requirements, an operation prompt suggesting cleaning is generated, informing the operator of the expected efficiency improvement. It is also explicitly stated that this predicted value is only for decision-making reference and is not used to modify the currently output actual cleaning efficiency value, ensuring that the authenticity and objectivity of the online detection results are not affected by human expectations.

[0117] This invention provides an implementation method for an online detection system for the impurity removal efficiency of a fluid iron separator, comprising: A high-frequency electromagnetic induction sensor is deployed in the inlet and outlet pipes of a fluid separator to measure the instantaneous content of magnetic impurities in the fluid in real time, generating inlet and outlet impurity content sequences. A miniature fluxgate sensor array is deployed on the surface of the magnetic medium inside the fluid separator to monitor the change of magnetic flux on the surface of the magnetic medium in real time and generate a magnetic flux decay curve. The edge computing gateway is connected to a high-frequency electromagnetic induction sensor and a miniature fluxgate sensor array, respectively, to receive the inlet impurity content sequence, the outlet impurity content sequence, and the flux decay curve. The edge computing gateway includes an apparent cleanup efficiency calculation unit, a magnetic medium adsorption saturation calculation unit, and a true cleanup efficiency fusion unit; The apparent impurity removal efficiency calculation unit is used to calculate the apparent impurity removal efficiency based on the inlet impurity content sequence and the outlet impurity content sequence. The magnetic medium adsorption saturation calculation unit is used to calculate the magnetic medium adsorption saturation based on the magnetic flux decay curve. The true impurity removal efficiency fusion unit is used to weight and fuse the apparent impurity removal efficiency by using the magnetic medium adsorption saturation as the adsorption effectiveness weight. When the magnetic medium adsorption saturation is lower than the preset saturation threshold, the true impurity removal efficiency fusion unit reduces the confidence of the apparent impurity removal efficiency and attenuates the apparent impurity removal efficiency according to the ratio of the true impurity removal efficiency to the apparent impurity removal efficiency multiplied by the magnetic medium adsorption saturation to the preset saturation threshold, so as to obtain the true impurity removal efficiency.

[0118] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. An on-line method for detecting the efficiency of a fluid de-ironer, characterized in that, The method comprises the following steps: deploying high-frequency electromagnetic induction sensors on the fluid de-ironing device inlet pipeline and the fluid de-ironing device outlet pipeline, and measuring the instantaneous content of magnetic impurities in the fluid in real time through the high-frequency electromagnetic induction sensors to obtain the inlet impurity content sequence and the outlet impurity content sequence; deploying a micro magnetic fluxgate sensor array on the surface of the magnetic medium inside the fluid de-ironing device, and monitoring the magnetic flux change of the surface of the magnetic medium in real time through the micro magnetic fluxgate sensor array to obtain the magnetic flux decay curve; the edge computing gateway receives the inlet impurity content sequence, the outlet impurity content sequence and the magnetic flux decay curve; the edge computing gateway calculates the apparent de-ironing efficiency according to the inlet impurity content sequence and the outlet impurity content sequence, and simultaneously calculates the magnetic medium adsorption saturation according to the magnetic flux decay curve; the edge computing gateway takes the magnetic medium adsorption saturation as the adsorption effectiveness weight to perform weighted fusion on the apparent de-ironing efficiency, when the magnetic medium adsorption saturation is lower than the preset saturation threshold, the confidence of the apparent de-ironing efficiency is reduced and the apparent de-ironing efficiency is attenuated according to the real de-ironing efficiency equaling to the apparent de-ironing efficiency multiplied by the ratio of the magnetic medium adsorption saturation to the preset saturation threshold, to obtain the real de-ironing efficiency.

2. The method of claim 1, wherein the method comprises: The step of calculating the magnetic medium adsorption saturation according to the magnetic flux decay curve comprises: the edge computing gateway performs first-order difference operation on the magnetic flux decay curve to obtain the magnetic flux decay rate curve; the edge computing gateway performs second-order difference operation on the magnetic flux decay rate curve to obtain the magnetic flux decay acceleration curve; the edge computing gateway searches for the time when the magnetic flux decay rate approaches the preset minimum rate threshold in the magnetic flux decay acceleration curve as the turning point when the adsorption layer on the surface of the magnetic medium changes from the stable adsorption state to the collapse critical state; the edge computing gateway takes the magnetic flux value corresponding to the turning point as the initial saturation magnetic flux, and takes the ratio of the current magnetic flux to the initial saturation magnetic flux as the magnetic medium adsorption saturation.

3. The method of claim 1, wherein the method further comprises: The step of calculating the magnetic medium adsorption saturation according to the magnetic flux decay curve comprises: the edge computing gateway obtains the magnetic flux values of multiple sampling points of the magnetic flux decay curve within a preset historical period; the edge computing gateway performs exponential function fitting on the magnetic flux values of the multiple sampling points to obtain a magnetic flux decay fitting function; the edge computing gateway extracts the decay time constant of the magnetic flux decay fitting function, and takes the ratio of the decay time constant to a preset reference time constant as the magnetic medium adsorption saturation, wherein the preset reference time constant is obtained by calibrating the adsorption experiment of the same specification of a brand-new fluid de-ironing device under standard working conditions.

4. The method of claim 1, wherein the method further comprises: The method further comprises the step of full flow field uniformity correction: deploying distributed magnetic sensors on the inlet section of the fluid de-ironing device internal flow channel, the middle section of the fluid de-ironing device internal flow channel and the outlet section of the fluid de-ironing device internal flow channel, and synchronously collecting magnetic impurity concentration data at different spatial positions through the distributed magnetic sensors; the edge computing gateway receives the magnetic impurity concentration data collected by the distributed magnetic sensors, and calculates the axial concentration gradient according to the magnetic impurity concentration data of the inlet section, the magnetic impurity concentration data of the middle section and the magnetic impurity concentration data of the outlet section. The edge computing gateway incorporates the axial concentration gradient as a flow field uniformity coefficient into the calculation of the actual impurity removal efficiency to obtain the final actual impurity removal efficiency.

5. The method of claim 4, wherein the method further comprises: The steps for incorporating the axial concentration gradient as a flow field uniformity coefficient into the calculation of the actual impurity removal efficiency include: The edge computing gateway acquires magnetic impurity concentration data for fine particle size range, medium particle size range, and coarse particle size range collected by distributed magnetic sensors. The edge computing gateway calculates the axial concentration gradients for the fine particle size range, the medium particle size range, and the coarse particle size range, respectively. The edge computing gateway performs a weighted summation of the axial concentration gradients of the fine particle size range, the medium particle size range, and the coarse particle size range according to preset weights to obtain the particle size weighted uniformity coefficient. When the particle size weighted uniformity coefficient exceeds the preset uniformity threshold, the edge computing gateway performs a second correction on the actual impurity removal efficiency. Finally, the actual impurity removal efficiency is equal to the actual impurity removal efficiency multiplied by the ratio of the preset uniformity threshold to the particle size weighted uniformity coefficient.

6. The method of claim 1, wherein the method further comprises: The step of calculating the apparent impurity removal efficiency based on the import and export impurity content sequences also includes a dynamic compensation step: A flow velocity sensor and a viscosity sensor are deployed in the inlet pipe of the fluid separator. The flow velocity sensor collects the fluid flow velocity in real time, and the viscosity sensor collects the fluid viscosity in real time. The edge computing gateway receives fluid velocity and fluid viscosity, and determines the dynamic compensation coefficient according to the ratio of fluid velocity to reference velocity and the ratio of fluid viscosity to reference viscosity in a linear function relationship. The dynamic compensation coefficient is equal to 1 plus the difference between the velocity compensation coefficient and the ratio of fluid velocity to reference velocity minus 1, plus the difference between the viscosity compensation coefficient and the ratio of fluid viscosity to reference viscosity minus 1. The edge computing gateway corrects the instantaneous magnetic impurity content in the inlet impurity content sequence and the instantaneous magnetic impurity content in the outlet impurity content sequence according to the dynamic compensation coefficient. The corrected instantaneous magnetic impurity content is equal to the original instantaneous magnetic impurity content multiplied by the dynamic compensation coefficient. The edge computing gateway uses the corrected inlet impurity content sequence and the corrected outlet impurity content sequence as input data to calculate the apparent impurity removal efficiency.

7. The method of claim 1, wherein the method further comprises: determining the amount of the contaminant in the fluid. The steps to obtain the import impurity content sequence and the export impurity content sequence include the sub-step of multi-frequency electromagnetic response identification: The high-frequency electromagnetic induction sensor alternately emits electromagnetic waves at a first detection frequency and a second detection frequency, wherein the first detection frequency is lower than the second detection frequency. The edge computing gateway receives a first response signal corresponding to a first detection frequency and a second response signal corresponding to a second detection frequency. According to the skin effect principle, when the first response signal is greater than or equal to the second response signal, the edge computing gateway takes the difference between the first response signal and the second response signal as the contribution value of weak magnetic impurities, and takes the remaining part of the second response signal after subtracting the contribution value of weak magnetic impurities as the content value of strong magnetic ferromagnetic impurities. When the first response signal is less than the second response signal, the contribution value of the weak magnetic impurities is set to zero, and the second response signal is directly used as the content value of the strong magnetic ferromagnetic impurities. The edge computing gateway regenerates the inlet and outlet impurity content sequences based on the content values ​​of strongly magnetic and ferromagnetic impurities, so that each content value in the inlet and outlet impurity content sequences corresponds to a strongly magnetic and ferromagnetic impurity that can be effectively removed by the fluid iron separator.

8. The method of claim 1, wherein the method further comprises: After obtaining the true impurity removal efficiency, the next step is trend prediction: The edge computing gateway obtains the actual noise removal efficiency calculated at multiple time points within a historical period, forming a historical sequence of actual noise removal efficiency. The edge computing gateway performs time-series analysis on the historical sequence of the actual noise removal efficiency to extract the decay rate of the actual noise removal efficiency; The edge computing gateway predicts the trend of real noise removal efficiency in the future based on the decay rate of real noise removal efficiency. When the predicted real noise removal efficiency will drop below the target efficiency threshold within a preset time window, an efficiency decay warning signal is generated.

9. The method of claim 1, wherein the method further comprises: determining the amount of the contaminant in the fluid. Following the steps to obtain the true cleaning efficiency, there is also a step to indicate the cleaning effect: The edge computing gateway obtains the historical cleaning records of the fluid iron separator. The historical cleaning records include the magnetic media adsorption saturation before each cleaning operation, the magnetic media adsorption saturation after each cleaning operation, and the actual impurity removal efficiency recovery value after each cleaning operation. The edge computing gateway establishes a cleaning effect prediction model based on historical cleaning records. The cleaning effect prediction model is used to characterize the relationship between the magnetic medium adsorption saturation and the actual cleanliness recovery value after cleaning. The edge computing gateway inputs the current magnetic medium adsorption saturation into the cleaning effect prediction model to obtain the predicted efficiency recovery value after the current cleaning operation, and generates cleaning timing suggestions and expected effect prompts based on the predicted efficiency recovery value, without correcting the current actual impurity removal efficiency.

10. An on-line detection system for the efficiency of a fluid de-ironer, characterized in that, include: A high-frequency electromagnetic induction sensor is deployed in the inlet and outlet pipes of a fluid separator to measure the instantaneous content of magnetic impurities in the fluid in real time, generating inlet and outlet impurity content sequences. A miniature fluxgate sensor array is deployed on the surface of the magnetic medium inside the fluid separator to monitor the change of magnetic flux on the surface of the magnetic medium in real time and generate a magnetic flux decay curve. The edge computing gateway is connected to a high-frequency electromagnetic induction sensor and a miniature fluxgate sensor array, respectively, to receive the inlet impurity content sequence, the outlet impurity content sequence, and the flux decay curve. The edge computing gateway includes an apparent cleanup efficiency calculation unit, a magnetic medium adsorption saturation calculation unit, and a true cleanup efficiency fusion unit; The apparent impurity removal efficiency calculation unit is used to calculate the apparent impurity removal efficiency based on the inlet impurity content sequence and the outlet impurity content sequence. The magnetic medium adsorption saturation calculation unit is used to calculate the magnetic medium adsorption saturation based on the magnetic flux decay curve. The true impurity removal efficiency fusion unit is used to weight and fuse the apparent impurity removal efficiency by using the magnetic medium adsorption saturation as the adsorption effectiveness weight. When the magnetic medium adsorption saturation is lower than the preset saturation threshold, the true impurity removal efficiency fusion unit reduces the confidence of the apparent impurity removal efficiency and attenuates the apparent impurity removal efficiency according to the ratio of the true impurity removal efficiency to the apparent impurity removal efficiency multiplied by the magnetic medium adsorption saturation to the preset saturation threshold, so as to obtain the true impurity removal efficiency.

Images