Abnormal data marking method and system for recovering elemental nickel from wastewater

Abstract

The present invention belongs to the field of recycling and data processing technology, and proposes a method and system for marking abnormal data of elemental nickel recovered from wastewater. Specifically, the method includes the following steps: first, initializing a recovery pool for recovering elemental nickel based on pulse current deposition technology; then collecting data from the recovery pool to obtain an adaptive input feature group, and performing an electronically controlled adaptive risk analysis based on the adaptive input feature group obtained in real time to form a risk-effective state value; finally, combining the electronically controlled adaptive risk analysis to mark abnormal data on the collected data. The defective position where the matching between the liquid phase environment and the nickel ion concentration collapses is marked, and the risk of the adaptability of the nickel ion concentration to the liquid phase environment when the sensitivity of the liquid phase environment is insufficient is effectively quantified based on a horizontal comparison of the defect position data of historical data. Therefore, reliable feedforward control parameters can be provided for the control parameter adjustment of the PLC and PID controllers, thereby improving the accuracy of the current density and reducing the risk of excessive loss of electrode materials and energy loss.

Description

Technical Field

[0001] The present invention belongs to the field of recycling and data processing technology, and in particular relates to a method and system for marking abnormal data of elemental nickel recovered from wastewater. Background Art

[0002] Many industrial production processes produce a large amount of nickel-containing wastewater, which poses a huge threat to the ecological environment and life safety. Among the methods for extracting and recovering elemental nickel from wastewater in industrial production today, membrane separation, electrochemical method and ion exchange method or a combination thereof are generally used. Among them, the recovery of elemental nickel in wastewater by a combination of ion exchange and electrodeposition is highly recognized for its high current efficiency and high recovery rate. For example, Patent No. CN201910878689.0 is a method and device for recovering elemental nickel from nickel-containing wastewater. During the implementation of this type of method, in order to achieve a preset deposition rate or effect, the control system of the electrodeposition equipment usually increases the current density; although a higher current density can accelerate the deposition process of elemental nickel, it will also increase energy consumption, resulting in excessive loss of electrode materials. In addition, metallic copper or iron or nickel have similarities in electrochemical properties, and therefore the risk of co-deposition problems will also increase. In order to scientifically control energy consumption and ensure deposition rate or deposition purity during the recovery of elemental nickel, pulse current electrodeposition technology is widely used. However, this method relies on real-time monitoring of the deposition liquid environment by PLC and PID controllers. Therefore, the controller control parameters such as pulse duration, frequency, and amplitude in the pulse current deposition technology need to be adjusted based on liquid environment monitoring. However, since the liquid environment of the input wastewater flow in the continuous flow electrodeposition system undergoes multiple pretreatments, the pH, temperature, and inflow nickel ion concentration in the wastewater are all unstable. The moving average method is generally used in this field to fit the liquid environment variables. However, this method is usually not sensitive enough to the liquid environment and is prone to ignoring the compatibility of nickel ion concentration with the liquid environment, resulting in hysteresis in the control parameter adjustment of PLC and PID controllers. When the nickel ion concentration drops significantly, the hysteresis of the system becomes stronger, which increases the unnecessary energy consumption of pulse current electrodeposition and forms energy loss. Therefore, there is an urgent need for a method for marking abnormal data for wastewater recovery of elemental nickel to monitor the compatibility of nickel ion concentration with the liquid environment to reduce the risk of energy loss. Summary of the Invention

[0003] The purpose of the present invention is to propose a method and system for marking abnormal data for recovering elemental nickel from wastewater, so as to solve one or more technical problems existing in the prior art and at least provide a beneficial option or create conditions.

[0004] In order to achieve the above object, according to one aspect of the present invention, a method for marking abnormal data of elemental nickel recovered from wastewater is provided, the method comprising the following steps:

[0005] S100, initializing a recovery pool for recovering elemental nickel based on pulsed current deposition technology;

[0006] S200, collecting data from the recycling pool to obtain an adapted input feature set;

[0007] S300, performing electronic control adaptation risk analysis based on the adaptation input feature set obtained in real time to form a risk effectiveness state value;

[0008] S400: Abnormal data is marked on the collected data in combination with the risk efficacy value.

[0009] Furthermore, in step S100, the method of initializing a recovery pool for recovering elemental nickel based on pulsed current deposition technology is: the recovery pool is used to recover elemental nickel from wastewater, and the recovery pool scene includes a nickel ion concentration measuring device, a temperature sensor, and a pH sensor, the nickel ion concentration measuring device is an inductively coupled plasma atomic emission spectrometer (ICP-OES) or an atomic absorption spectrometer (AAS), the temperature sensor is an infrared temperature sensor or a platinum resistance temperature detector, and the pH sensor is an optical pH sensor or a solid-state pH sensor, wherein the solid-state pH sensor is an ion sensitive field effect transistor (ISFET); the pretreated wastewater is input into the recovery pool through a transmission pipeline; and the nickel ion concentration measuring device, the temperature sensor, and the pH sensor are arranged in the transmission pipeline.

[0010] The pretreatment of wastewater includes filtration of suspended solids and particulate matter, adsorption of organic matter and grease, adsorption with cation exchange resin, and impurity removal with anion exchange resin.

[0011] Furthermore, in step S200, data is collected from the recovery tank to obtain an adapted input feature set by: recording the sampling interval as IG, IG∈[5,10] seconds, and the measurement point window value at a moment is the IG period in the reverse time direction of the moment; marking a moment as a measurement point every IG, and recording the liquid phase residence time as TS, where the liquid phase residence time refers to the time that the liquid stays in the system in the continuous flow electrodeposition system;

[0012] The nickel ion concentration is measured in real time by a nickel ion concentration measuring device and recorded as the concentration value. The average concentration value within the measurement point interval is the measurement point window value, that is, the average value of each concentration value obtained from a moment to the previous IG time period is the measurement point window value;

[0013] The difference between the concentration value at any moment and its measurement point window value is recorded as the first mean difference; the mean of all concentration values ​​from a moment to the previous TS is recorded as the dwell window value; the difference between the concentration value at the current moment and the dwell window value is recorded as the second mean difference, and the difference between the first mean difference and the second mean difference at the current moment is the sub-concentration tilt. For a measurement point, the maximum value of the sub-concentration tilt obtained at each moment between the measurement point and the first measurement point in the reverse time direction is recorded as the concentration tilt of the measurement point;

[0014] The temperature and pH values ​​are measured in real time by a temperature sensor and a pH sensor, respectively. The temperature and pH values ​​are combined into a binary pair and recorded as a liquid phase array. The average value of each liquid phase array within the TS time period is obtained and recorded as the reference liquid phase array. The Euclidean distance between the liquid phase array obtained at any measuring point and the reference liquid phase array is calculated and recorded as the sub-liquid phase distance. The maximum value of several sub-liquid phase distances obtained within a sampling interval is recorded as the liquid phase distance. An adapted input feature group is obtained at each measuring point, where the adapted input feature group is a binary pair formed by the liquid phase distance and the concentration tilt.

[0015] Traditional static threshold-based methods have limitations when dealing with dynamically changing wastewater treatment processes. For nickel recovery, potential features include flow rate, pH, ORP redox potential, temperature, conductivity, current / voltage in electrochemical methods, and the concentration of the chemical substances used. Feature engineering is used to screen these features based on rolling statistics of defect location data, which is a match between the labeled liquid environment and the nickel ion concentration. A subset of available features is dynamically selected to determine the preferred liquid phase distance and concentration tilt for this method. This feature engineering is a sparse learning method based on the self-attention mechanism and LASSO regression. Selecting the adapted input feature set for adaptive feature selection can reduce noise and irrelevant information that may cause false positives, and shorten the execution process inference time to reduce the memory requirements for storing and processing data.

[0016] The liquidus distance reflects the stability of the liquidus environment or the degree of local perturbation. Since the electrodeposition process is very sensitive to temperature and pH, the migration rate of metal ions near the electrode interface and the deposition morphology are affected. The liquidus distance can promptly detect environmental changes that may lead to abnormal electrodeposition. A larger liquidus distance indicates a more severe liquidus environment fluctuation. In a continuous flow electrodeposition system, severe liquidus environment fluctuations may affect the deposition efficiency and quality of nickel ions and increase the risk of co-deposition. Therefore, the liquidus distance is used as an indicator to capture the risk of liquidus environment mutations. The concentration tilt reflects the trend of nickel ion concentration. In a continuous flow electrodeposition system, the nickel ion concentration of the input wastewater is a key environmental indicator that can directly point to the control parameters of the electrodeposition equipment. However, the uncertainty of the nickel ion concentration trend and the lack of adaptation to the uncertainty of the liquidus environment often lead to errors in the adjustment of the control parameters of the electrodeposition equipment, resulting in unnecessary energy consumption during long-term operation. Therefore, energy loss risk analysis is performed on the feature engineering results. From the perspective of control theory, the liquid phase distance and concentration tilt correspond to the amplitude term of the input disturbance and the derivative term of the target variable, respectively. The two together constitute the main factors affecting the controller response delay and adjustment error.

[0017] The real-time measurement frequency of the nickel ion concentration measuring device, temperature sensor, and pH sensor is the same, which is less than the sampling interval. The sampling interval is at least 5 times the real-time measurement frequency, that is, at least 5 real-time measurements are performed and measured values ​​are formed within each sampling interval. When the second mean difference at a measuring point is 0, the sub-concentration tilt is not calculated for that measuring point.

[0018] Furthermore, in step S300, the method for performing electronically controlled adaptive risk analysis to form a risk effectiveness value based on the adaptive input feature group obtained in real time is as follows: a time period is set as a deposition monitoring interval NIMD, and its value range is set to NIMD∈[5, 10] minutes. Within the current NIMD period, all the liquid phase distances and concentration inclinations corresponding to the adaptive input feature groups are respectively formed into sequences, and recorded as liquid phase distance sequence ALP and concentration inclination sequence ALC; j1 is used as the serial number of the measuring point, and in the liquid phase distance sequence, the ratio of the deviation between any element and the element in the liquid phase distance sequence is recorded as the phase deviation value of the measuring point corresponding to the element; the phase deviation value at each measuring point is normalized to form a deviation modulus sequence MdsL, and the phase self-bias feature PSd of the j1-th measuring point can be calculated based on the liquid phase distance sequence and the deviation modulus sequence. j1 , which is calculated as follows: PSd j1=exp(ALP(j1)×MdsL(j1)); where exp() represents an exponential function with base e, ALP(j1) and MdsL(j1) represent the element corresponding to the j1th measuring point in the liquid phase distance sequence and the offset modulus sequence, respectively; the ratio of the element under any measuring point in the concentration tilt sequence to the extreme value of the concentration tilt sequence is recorded as the tilt value TdV of the measuring point;

[0019] The phase self-bias characteristic is designed to evaluate the degree of deviation of the liquid phase distance at each measurement point relative to the overall sequence and to weight the liquid phase distance. An exponential function nonlinearly converts this weighted deviation into a final anomaly score. This method is intended to highlight moments when significant and abnormal changes occur in the liquid phase environment, making the score more sensitive to larger deviations, indicating the instability of the electrodeposition process.

[0020] The corresponding measuring points of each element in the sequence ALC whose value is greater than zero are recorded as increasing points, otherwise they are recorded as decreasing points. Increasing points and decreasing points are regarded as two subcategories of the class point classification; each measuring point is traversed in reverse chronological order, and the measuring point being traversed is regarded as the traversing measuring point, and the measuring point that has been traversed is regarded as the historical traversing measuring point. The proportion of measuring points in each historical traversing measuring point that are classified as the same class point as the traversing measuring point is recorded as the homology ratio SRt. If the occurrence time of the traversing measuring point is before 0.2NIMD of the current time, the concentration tilt Ct0 of the traversing measuring point is down-weighted fitting, that is, the concentration tilt in the sequence ALC is updated to Ct = SRt×Ct0 / (1-SRt); then the risk effect state value is the product of the relative self-bias characteristic and the percentile value of the current measuring point in the normalized sequence ALC. The phase autobias characteristic refers to the phase autobias characteristic value of the current measuring point. The sequence ALC after weight adjustment is normalized. The percentile value of the corresponding value of the current measuring point after normalization is the weighted percentile. The product of the weighted percentile and the phase autobias characteristic of the current measuring point is the risk effect value.

[0021] The change in the trend of nickel ion concentration during wastewater electroplating will cause a sudden increase in the phase autobias, resulting in a significant deviation in the risk-effective state value compared with the normal state. However, this process shows that the hysteresis problem in the recovery process efficiency does not directly point to the problem of insufficient adaptation of electroplating to the liquid phase environment, but is still directly related to other factors such as ion exchange resin saturation or system leakage. Therefore, it is necessary to screen this adaptability through the percentile value of the current measurement point sequence ALC. The larger the percentile value of the current measurement point, the stronger the directionality of the adaptability of electroplating to the liquid phase environment. This is because the percentile value of the measurement point in the normalized ALC sequence is consistent with the typical fluctuations observed during normal operation. The higher the consistency, the stronger the applicability of its dynamic judgment rules, thereby reducing the risk of misjudgment caused by non-dynamic faults.

[0022] Since the calculation of the above-mentioned risk effective state value is too limited to the data of the inclination distance value, even if the data characteristics of the compatibility risk of nickel ion concentration and liquid environment when the sensitivity of the liquid environment is insufficient are effectively identified, this calculation method makes the screening of liquid phase distance and concentration inclination weaken the data difference between each moment, resulting in a phenomenon of excessive focus on global information and neglect of local characteristics, which makes the quantification result have the risk of timeliness loss; however, the existing technology cannot solve the problem of this timeliness loss risk. In order to make the use of risk effective state value more adaptable and eliminate the phenomenon of insufficient independence between the data, the present invention proposes a more preferred solution.

[0023] Preferably, in step S300, the method for performing electronically controlled adaptive risk analysis based on the adaptive input feature group obtained in real time to form a risk effectiveness state value is as follows: setting a time period as a deposition monitoring interval NIMD_a, setting its value range to NIMD_a∈[5,20] minutes, and extracting the concentration inclinations of each adaptive input feature group within the most recent NIMD_a time period to form a concentration inclination sequence;

[0024] When the concentration tilt Ct of a measuring point is greater than 0 and the concentration tilt is greater than or equal to the upper quartile of the concentration tilt sequence, the measuring point is recorded as a high-tilt site. When the concentration tilt Ct corresponding to a measuring point is less than 0 and the concentration tilt is less than or equal to the lower quartile of the concentration tilt sequence, the measuring point is recorded as a low-tilt site. The average values ​​of the liquid phase distances corresponding to each high-tilt site and each low-tilt site are recorded as the high-tilt threshold HtsV and the low-tilt threshold LtsV, respectively.

[0025] If there is at least one low-dip site in the time sequence search from a measuring point, and at least one high-dip site in the reverse time sequence search, then the measuring point is recorded as the separation point; the liquid phase distance corresponding to the first low-dip site in the time sequence search from the separation point is recorded as the low phase distance Hpd, and the liquid phase distance corresponding to the first high-dip site in the reverse time sequence search from the separation point is recorded as the high phase distance Lpd;

[0026] If a bidirectional search at any measuring point fails to obtain a low-inclination point or a high-inclination point, no subsequent operation will be performed on the elements under this measuring point, that is, the start and end data of the adapted input feature group will be ignored.

[0027] The absolute value of the ratio of the high-order phase distance to the low-order phase distance of any separation point is recorded as the distance deviation DoL of the separation point, and the liquid phase distances corresponding to each separation point are written into a sequence recorded as the separation sequence ANs Lp , write the concentration tilt corresponding to each precipitation point into a sequence and record it as the decantation sequence ANs Ct , and ANs Lp [.] represents the elements in the parsed sequence, with ANs Ct[.] represents the elements in the decantation sequence, i1 is used as the sequence number of the decantation point, and the effective phase state PteS of the i1th decantation point is calculated based on the low-order phase distance and the high-order phase distance:

[0028] ;

[0029] Among them ANs Lp [i1] ​​represents the i1th element in the distance analysis sequence, DoL i1 Represents the distance deviation of the i1th distance analysis point in the distance analysis sequence, Hpd i1 With Lpd i1 Respectively represent the high and low distances of the i1th distance analysis point, exp() represents the exponential function with e as the base; the ratio of the average absolute deviation of any element in the distance analysis sequence to the sequence is recorded as the skewness Opv of the element;

[0030] Take i2 as the serial number of the measuring point, and calculate the risk effect state value Rpv of the i2th measuring point according to the phase skewness and phase distance effect state i2 , which is calculated as follows:

[0031] ;

[0032] Among them, i3 is the cumulative variable, NIS is the number of distance points, PteS i1 is the effective state of the i1th element in the distance analysis sequence, ANs Lp [i1] ​​and ANs Ct [i1] ​​is the i1th element in the distance and decantation sequences, ALC[i2] represents the i2th element in the concentration tilt sequence, Opv i3 is the i3th element in the distance analysis sequence, avg<> is the average function, hs{} is the harmonic mean function, function In the equation, the value range of i1 is i1∈[1,NIS].

[0033] Beneficial effects: Since the risk-effective state value is calculated based on the liquid phase state and the change in nickel ion concentration, it can accurately mark the defect location where the matching collapse of the liquid phase environment and nickel ion concentration. Based on the horizontal comparison of the defect location data of historical data, the adaptability risk of nickel ion concentration and liquid phase environment when the sensitivity of the liquid phase environment is insufficient is effectively quantified. Therefore, it can provide reliable feedforward control parameters for the control parameter adjustment of PLC and PID controllers, thereby improving the accuracy of current density and reducing the risk of excessive loss of electrode materials and energy loss.

[0034] Furthermore, in step S400, the method for marking abnormal data for the collected data in combination with the risk effective state value is as follows: each time the risk effective state value is calculated, the corresponding risk effective state value is obtained at each measuring point within the current deposition monitoring interval; if a measuring point obtains the risk effective state value for the first time, the risk effective state value is recorded as the effective state base value of the measuring point, and a deposition setting group is obtained in real time, and the effective state base value and the deposition setting group are formed into a binary group as an adjustment identification group of the measuring point; wherein the deposition setting group is a multivariate array composed of measured values ​​corresponding to several deposition options, and the measured values ​​corresponding to the deposition options are obtained through a PID controller or a PLC controller, and the deposition options include: one or more of current, current density, voltage and electrode spacing; if a measuring point does not obtain the risk effective state value for the first time, the risk effective state value calculated for the measuring point is recorded as an effective state derivative value; among the various effective state derivative values ​​of any measuring point, the proportion of the effective state derivative values ​​with values ​​smaller than the effective state base value is taken as the effective state overflow ratio of the measuring point;

[0035] Set a time period as the monitoring period (TSS), with a value range of TSS∈[5,10] minutes. Define the measurement point whose effective state overflow ratio exceeds 50% in the TSS period before the current moment as the effective state reference measurement point at the current moment. Perform a box plot analysis on the adjustment identification group of the current measurement point and each effective state reference measurement point to identify outliers. If the current measurement point is an outlier, a warning is issued that the current electrodeposition adjustment system performance has a hysteresis risk, and the measured values ​​of each deposition option obtained at that moment are marked as abnormal data. The current measurement point is the measurement point closest to the current moment.

[0036] Whether the measurement point is an abnormal point is used as an abnormal label. When the data collection volume is sufficient, an abnormality prediction model is constructed through a machine learning algorithm. The abnormal label is used as the prediction label of the supervised learning model. The measured value of the deposition option, the pH value of the adapted input feature group, the temperature value and several values ​​of the nickel ion concentration are used as training features to train the big data model. The big data model can be any of the supervised learning models. The prediction model obtained through training is used to adjust the control parameters of the PLC and PID controller, thereby improving the accuracy and timeliness of current density control.

[0037] Preferably, all undefined variables in the present invention, if not clearly defined, can be manually set thresholds. The description of the current period in the present invention is defined as a period of time in the reverse direction of the current moment.

[0038] The present invention also provides an abnormal data marking system for recovering elemental nickel from wastewater. The abnormal data marking system for recovering elemental nickel from wastewater includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the abnormal data marking method for recovering elemental nickel from wastewater are implemented. The abnormal data marking system for recovering elemental nickel from wastewater can be run on computing devices such as desktop computers, laptop computers, PDAs, and cloud data centers. The executable system may include, but is not limited to, a processor, a memory, and a server cluster. The processor executes the computer program to run in the following system units:

[0039] A recovery pool initialization unit, used to initialize a recovery pool for recovering elemental nickel based on pulsed current deposition technology;

[0040] A data acquisition unit is used to collect data from the recycling pool to obtain an adapted input feature set;

[0041] The electronic control adaptation risk analysis unit is used to perform electronic control adaptation risk analysis based on the adaptation input feature set obtained in real time to form a risk effect state value;

[0042] The abnormal data marking unit is used to mark the collected data as abnormal data in combination with the risk efficacy value.

[0043] The beneficial effects of the present invention are as follows: the present invention provides a method and system for marking abnormal data of elemental nickel recovered from wastewater, which identifies the defective position where the matching between the liquid phase environment and the nickel ion concentration collapses based on the data of the liquid phase state and the nickel ion concentration change, and effectively quantifies the risk of the compatibility between the nickel ion concentration and the liquid phase environment when the sensitivity of the liquid phase environment is insufficient based on the horizontal comparison of the defect position data of historical data. Therefore, it can provide reliable feedforward control parameters for the control parameter adjustment of the PLC and PID controller, thereby improving the accuracy of the current density and reducing the risk of excessive loss of electrode materials and energy loss. The frequency of occurrence of this type of abnormal data in wastewater recovery scenarios with a low degree of standardization can more efficiently identify the hysteresis problem of the current density control, making the utilization rate of energy consumption and production resources higher, thereby improving the stability of the wastewater recovery system. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The above and other features of the present invention will become more apparent through a detailed description of the embodiments shown in conjunction with the accompanying drawings. In the drawings of the present invention, the same reference numerals represent the same or similar elements. Obviously, the drawings described below are only some embodiments of the present invention. It is possible for a person skilled in the art to derive other drawings based on these drawings without inventive effort. In the drawings:

[0045] Figure 1 Shown is a flow chart of a method for marking abnormal data for recovering elemental nickel from wastewater;

[0046] Figure 2 Shown is a structural diagram of an abnormal data marking system for recovering elemental nickel from wastewater. DETAILED DESCRIPTION

[0047] The following will be combined with the embodiments and drawings to clearly and completely describe the concept, specific structure and technical effects of the present invention so as to fully understand the purpose, scheme and effect of the present invention. It should be noted that the embodiments and features in the embodiments of this application can be combined with each other unless there is a conflict.

[0048] Example 1: Figure 1 The figure shows a flow chart of a method for marking abnormal data of elemental nickel recovered from wastewater. Figure 1 To illustrate a method for marking abnormal data of elemental nickel recovered from wastewater according to an embodiment of the present invention, the method comprises the following steps:

[0049] S100, initializing a recovery pool for recovering elemental nickel based on pulsed current deposition technology;

[0050] S200, collecting data from the recycling pool to obtain an adapted input feature set;

[0051] S300, performing electronic control adaptation risk analysis based on the adaptation input feature set obtained in real time to form a risk effectiveness state value;

[0052] S400: Abnormal data is marked on the collected data in combination with the risk efficacy value.

[0053] Furthermore, in step S100, the method of initializing a recovery pool for recovering elemental nickel based on pulsed current deposition technology is: the recovery pool is used to recover elemental nickel from wastewater, and the recovery pool scene includes a nickel ion concentration measuring device, a temperature sensor, and a pH sensor, the nickel ion concentration measuring device is an inductively coupled plasma atomic emission spectrometer or an atomic absorption spectrometer, the temperature sensor is an infrared temperature sensor or a platinum resistance temperature detector, and the pH sensor is an optical pH sensor or a solid-state pH sensor, wherein the solid-state pH sensor is an ion-sensitive field-effect transistor (ISFET); the pretreated wastewater is input into the recovery pool through a transmission pipeline; and the nickel ion concentration measuring device, the temperature sensor, and the pH sensor are arranged in the transmission pipeline.

[0054] Furthermore, in step S200, data is collected from the recovery pool to obtain an adapted input feature set in the following manner: the sampling interval is recorded as IG, which is 10 seconds; a moment is marked as a measuring point every IG, and the liquid phase residence time is recorded as TS; the concentration of nickel ions is measured in real time by a nickel ion concentration measuring device and recorded as a concentration value, the average value of each concentration value obtained from a moment to the previous IG is the measuring point window value, and the difference between the concentration value at the current moment and the measuring point window value is recorded as the first mean difference; the average value of each concentration value from a moment to the previous TS is the residence window value; the difference between the concentration value at the current moment and the residence window value is recorded as the second mean difference, and the difference between the first mean difference at the current moment and the second mean difference is the sub-concentration tilt; for a measuring point, the maximum value of the sub-concentration tilt obtained at each moment between the measuring point and the first measuring point in the reverse time direction is The value is recorded as the concentration tilt of the measuring point; the temperature value and pH value are respectively measured in real time by the temperature sensor and the pH sensor, and the temperature value and pH value are formed into a binary group and recorded as the liquid phase array. The average value of each liquid phase array in the TS time period is obtained and recorded as the reference liquid phase array; the Euclidean distance between the liquid phase array obtained at any measuring point and the reference liquid phase array is calculated and recorded as the sub-liquid phase distance; the maximum value of several sub-liquid phase distances obtained within a sampling interval is recorded as the liquid phase distance; each measuring point obtains an adaptive input feature group, where the adaptive input feature group is a binary group formed by the liquid phase distance and the concentration tilt.

[0055] Furthermore, in step S300, the method for performing electronically controlled adaptive risk analysis to form a risk effectiveness value based on the adaptive input feature group obtained in real time is as follows: a time period is set as a deposition monitoring interval NIMD, and its value range is set to 10 minutes of NIMD value. Within the current NIMD period, all the liquid phase distances and concentration inclinations corresponding to the adaptive input feature groups are respectively formed into sequences, and recorded as liquid phase distance sequence ALP and concentration inclination sequence ALC; j1 is used as the serial number of the measuring point, and in the liquid phase distance sequence, the ratio of the deviation of any element and the element in the liquid phase distance sequence is recorded as the phase deviation value of the measuring point corresponding to the element; the phase deviation value at each measuring point is normalized to form a deviation modulus sequence MdsL, and the phase self-bias feature PSd of the j1-th measuring point can be calculated based on the liquid phase distance sequence and the deviation modulus sequence. j1 , which is calculated as follows: PSd j1 =exp(ALP(j1)×MdsL(j1)); where exp() represents an exponential function with base e, ALP(j1) and MdsL(j1) represent the element corresponding to the j1th measuring point in the liquid phase distance sequence and the offset modulus sequence, respectively; the ratio of the element under any measuring point in the concentration tilt sequence to the extreme value of the concentration tilt sequence is recorded as the tilt value TdV of the measuring point;

[0056] The corresponding measuring points of each element in the sequence ALC whose value is greater than zero are recorded as increasing points, otherwise they are recorded as decreasing points. Increasing points and decreasing points are regarded as two subcategories of the class point classification; each measuring point is traversed in reverse chronological order, and the measuring point being traversed is regarded as the traversing measuring point, and the measuring point that has been traversed is regarded as the historical traversing measuring point. The proportion of measuring points in each historical traversing measuring point that are classified as the same class point as the traversing measuring point is recorded as the homology ratio SRt. If the occurrence time of the traversing measuring point is before 0.2NIMD of the current time, the concentration tilt Ct0 of the traversing measuring point is down-weighted fitting, that is, the concentration tilt in the sequence ALC is updated to Ct = SRt×Ct0 / (1-SRt); then the risk effect state value is the product of the relative self-bias characteristic and the percentile value of the current measuring point in the normalized sequence ALC.

[0057] Furthermore, in step S400, the method for marking abnormal data for the collected data in combination with the risk effective state value is as follows: each time the risk effective state value is calculated, the corresponding risk effective state value is obtained at each measuring point within the current deposition monitoring interval; if a measuring point obtains the risk effective state value for the first time, the risk effective state value is recorded as the effective state base value of the measuring point, and a deposition setting group is obtained in real time, and the effective state base value and the deposition setting group are formed into a binary group as an adjustment identification group of the measuring point; wherein the deposition setting group is a multivariate array consisting of actual measured values ​​corresponding to a number of deposition options, and the actual measured values ​​corresponding to the deposition options are obtained through a PID controller or a PLC controller, and the deposition options include: current and current density; if a measuring point does not obtain the risk effective state value for the first time, the risk effective state value calculated for the measuring point is recorded as an effective state derivative value; among the various effective state derivative values ​​of any measuring point, the proportion of effective state derivative values ​​with values ​​smaller than the effective state base value is taken as the effective state overflow ratio of the measuring point;

[0058] Set a time period as the monitoring section TSS value of 10 minutes; define the measuring point with an effective state overflow ratio exceeding 50% in the TSS period before the current moment as the effective state reference measuring point at the current moment; perform box plot anomaly identification on the adjustment identification group of the current measuring point and each effective state reference measuring point. If the current measuring point is an anomaly point, it will warn that there is a hysteresis risk in the performance of the current electro-deposition adjustment system, and mark the actual measured values ​​of each deposition option obtained at that moment as abnormal data.

[0059] Example 2: Example 2 uses the same marking process as Example 1, except that, in step S300, the method for performing electronic control adaptation risk analysis based on the adaptation input feature set obtained in real time to form a risk efficacy state value is:

[0060] Set a time period as the sedimentation monitoring interval NIMD_a, set its value to 10 minutes, and extract the concentration tilt of each adapted input feature group within the most recent NIMD_a time period to form a concentration tilt sequence;

[0061] When the concentration tilt Ct of a measuring point is greater than 0 and the concentration tilt is greater than or equal to the upper quartile of the concentration tilt sequence, the measuring point is recorded as a high-tilt site. When the concentration tilt Ct of a measuring point is less than 0 and the concentration tilt is less than or equal to the lower quartile of the concentration tilt sequence, the measuring point is recorded as a low-tilt site. The average values ​​of the liquid phase distances corresponding to each high-tilt site and each low-tilt site are recorded as the high-tilt threshold HtsV and the low-tilt threshold LtsV, respectively. During the current NIMD_a period, if there is at least one low-tilt site in the chronological search from a measuring point and at least one high-tilt site in the reverse chronological search, the measuring point is recorded as a separation point. The liquid phase distance corresponding to the first low-tilt site in the chronological search from a separation point is recorded as the low-phase distance Hpd, and the liquid phase distance corresponding to the first high-tilt site in the reverse chronological search from a separation point is recorded as the high-phase distance Lpd.

[0062] The absolute value of the ratio of the high-order phase distance to the low-order phase distance of any separation point is recorded as the distance deviation DoL of the separation point, and the liquid phase distances corresponding to each separation point are written into a sequence recorded as the separation sequence ANs Lp , write the concentration tilt corresponding to each precipitation point into a sequence and record it as the decantation sequence ANs Ct , taking i1 as the sequence number of the distance analysis point, calculate the effective distance state PteS of the i1th distance analysis point according to the low-order distance and the high-order distance:

[0063] ;

[0064] Among them ANs Lp [i1] ​​represents the i1th element in the distance analysis sequence, DoL i1 Represents the distance deviation of the i1th distance analysis point in the distance analysis sequence, Hpd i1 With Lpd i1 Respectively represent the high and low distances of the i1th distance analysis point, exp() represents the exponential function with e as the base; the ratio of the average absolute deviation of any element in the distance analysis sequence to the sequence is recorded as the skewness Opv of the element;

[0065] Take i2 as the serial number of the measuring point, and calculate the risk effect state value Rpv of the i2th measuring point according to the phase skewness and phase distance effect state i2 , which is calculated as follows:

[0066] ;

[0067] Among them, i3 is the cumulative variable, NIS is the number of distance points, PteS i1 is the effective state of the i1th element in the distance analysis sequence, ANs Lp [i1] ​​and ANs Ct[i1] ​​is the i1th element in the distance and decantation sequences, ALC[i2] represents the i2th element in the concentration tilt sequence, Opv i3 is the i3th element in the distance analysis sequence, avg<> is the average function, hs{} is the harmonic mean function, function In the equation, the value range of i1 is i1∈[1,NIS].

[0068] Comparative Example:

[0069] The method for identifying abnormal data based on moving average trend judgment, which is commonly used in traditional industrial wastewater electrodeposition control, uses a sliding window to obtain the average values ​​of liquid phase temperature, pH value, and nickel ion concentration and compare them with the threshold value to determine whether the system has entered an abnormal state. The specific settings are as follows: a 10-minute sliding window is used to perform mean fitting on each liquid phase parameter. The abnormal point judgment logic is: if the current measured value exceeds the window mean ±2 times the standard deviation, it is judged to be abnormal. The controller adjustment only corrects the set parameters based on the current liquid phase value, and does not introduce risk-effective state values ​​or homology ratio models. The data acquisition period is consistent with Examples 1 and 2, with a sampling interval of 10 seconds and a liquid phase residence time of 20 seconds. The test period is the same as that of Examples 1 and 2, both of which are 30 days.

[0070]

[0071] Table 1 Comparison of electrodeposition process performance under different labeling methods

[0072] As shown in Table 1, the comparative example has no obvious advantages in terms of algorithm latency and resource consumption. However, it performs poorly in key performance indicators such as anomaly identification accuracy, energy consumption reduction, material loss control, and output purity. In particular, it has obvious shortcomings in the ability to quickly identify violent fluctuations in the liquid phase and sudden changes in the adaptability of nickel ion concentration, resulting in energy waste and accelerated material aging. However, Examples 1 and 2 proposed by the present method significantly improve the timeliness and directionality of anomaly identification by introducing risk-effectiveness value modeling and homology ratio fitting mechanisms, achieving more precise regulation, thereby optimizing the operating efficiency and stability of the wastewater recovery system as a whole.

[0073] In practical applications, although both Example 1 and Example 2 use the risk efficacy value as the core discrimination indicator, their applicable scenarios are different, and strategy selection needs to be combined with the specific manifestations of liquid phase environment fluctuations.

[0074] Example 1 is applicable to the construction of a full-scale anomaly recognition model for each measuring point in a continuous flow wastewater recovery system. The model does not rely on the structural connection between the data, but takes the liquid phase distance and concentration tilt of each measuring point as input, and nonlinearly maps the relative relationship between the current liquid phase environment and the global fluctuation through an exponential function, thereby capturing the macro stability offset. This method can stably output the adaptability signal of the electrodeposition environment at each moment under the condition that the overall fluctuation of the system is relatively stable and there is no significant mutation. It is suitable for real-time global monitoring and ensures that the control system has feedforward correction capabilities. While Example 2 embodies the response chain of the liquid phase perturbation gradually transitioning from the excited state to the buffer state during the electrodeposition process at the microscopic level, it is more suitable for the working condition where the liquid phase environment presents an obvious oscillation mode or periodic deviation behavior on the time axis, that is, the liquid phase condition and the nickel ion concentration alternately show a trend of rapid rise and fall in a short time. Such features are often caused by the transient feedback mechanism caused by uneven acid and alkali addition in the raw wastewater, pretreatment adsorption imbalance or temperature disturbance. In this oscillation mode, if Example 1 is still used for point analysis, it is easy to have insufficient local sensitivity. Therefore, Example 1 of the present invention provides a basic risk modeling method, which is suitable for continuous monitoring throughout the entire period; Example 2 is a preferred implementation method, which is only used in areas where liquid phase mutations are obvious. The two methods can be replaced with each other according to the system's preset switching logic and do not constitute a technical implementation obstacle.

[0075] The embodiment of the present invention provides a wastewater recovery elemental nickel abnormal data marking system, such as Figure 2 Shown is a structural diagram of an abnormal data marking system for recovering elemental nickel from wastewater according to the present invention. The abnormal data marking system for recovering elemental nickel from wastewater in this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the above-mentioned embodiment of the abnormal data marking system for recovering elemental nickel from wastewater are implemented.

[0076] The system includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to run in the following units of the system:

[0077] A recovery pool initialization unit, used to initialize a recovery pool for recovering elemental nickel based on pulsed current deposition technology;

[0078] A data acquisition unit is used to collect data from the recycling pool to obtain an adapted input feature set;

[0079] The electronic control adaptation risk analysis unit is used to perform electronic control adaptation risk analysis based on the adaptation input feature set obtained in real time to form a risk effect state value;

[0080] The abnormal data marking unit is used to mark the collected data as abnormal data in combination with the risk efficacy value.

[0081] The abnormal data marking system for recycling elemental nickel from wastewater can be run on computing devices such as desktop computers, laptop computers, PDAs, and cloud servers. The system for recycling elemental nickel from wastewater can include, but is not limited to, a processor and a memory. Those skilled in the art will understand that the example is merely an example of an abnormal data marking system for recycling elemental nickel from wastewater, and does not constitute a limitation on an abnormal data marking system for recycling elemental nickel from wastewater. The system can include more or fewer components than the example, or a combination of certain components, or different components. For example, the abnormal data marking system for recycling elemental nickel from wastewater can also include input and output devices, network access devices, buses, etc.

[0082] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the operation system of the abnormal data marking system for recovering elemental nickel from wastewater, and utilizes various interfaces and lines to connect various parts of the entire operation system of the abnormal data marking system for recovering elemental nickel from wastewater.

[0083] The memory can be used to store the computer programs and / or modules. The processor implements the various functions of the abnormal data marking system for recovering elemental nickel from wastewater by running or executing the computer programs and / or modules stored in the memory and accessing the data stored in the memory. The memory may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as a sound playback function or an image playback function); the data storage area may store data generated based on the use of the mobile phone (such as audio data and a phone book). Furthermore, the memory may include high-speed random access memory (RAM) and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0084] Although the present invention has been described in considerable detail and with particularity with respect to several embodiments, it is not intended to limit the present invention to any of these details or embodiments or any particular embodiment, so as to effectively encompass the intended scope of the present invention. In addition, the present invention has been described above with respect to embodiments foreseen by the inventors for the purpose of providing a useful description, and those insubstantial modifications of the present invention that are not currently foreseen may still represent equivalent modifications of the present invention.

Claims

1. A method for marking abnormal data for recovering elemental nickel from wastewater, characterized in that: The method comprises the following steps: S100, initializing a recovery pool for recovering elemental nickel based on pulsed current deposition technology; S200, collecting data from the recycling pool to obtain an adapted input feature set; S300, performing electronic control adaptation risk analysis based on the adaptation input feature set obtained in real time to form a risk effectiveness state value; S400, marking abnormal data on the collected data in combination with the risk efficacy value; Specifically, step S200 includes collecting data from the recovery pool to obtain an adaptive input feature set. The method is as follows: the liquid phase residence time is recorded as TS; the mean of the concentration values ​​within the measurement point interval at a time is subtracted from the mean of the concentration values ​​within TS to form a first mean difference and a second mean difference; the maximum value of the difference between the first mean difference and the second mean difference within the measurement point interval is the concentration tilt; the temperature value and the pH value are used to form a liquid phase array, and the Euclidean distance between the liquid phase array at a time and the average value of each liquid phase array within the previous TS time period is calculated. The maximum value of the obtained Euclidean distance is the liquid phase distance, and the liquid phase distances of the measurement points and the concentration tilt constitute the adaptive input feature set. Step S300 specifically includes: presetting a monitoring interval and forming a sequence ALP and a sequence ALC of all liquid phase distances and concentration tilts therein, respectively, to construct an exponential function to obtain a phase autobias characteristic; using a value of 0 as a classification threshold, dividing the measurement points into increasing and decreasing points according to the sequence ALC; traversing each measurement point in reverse chronological order at the current measurement point, taking the measurement point being traversed as the current measurement point, and taking the measurement points that have been traversed as historical traversed measurement points; the proportion of measurement points in each historical traversed measurement point that are classified as the same class as the current measurement point is the homology ratio SRt; performing a weighted fitting on the early ALC data Ct0 using the homology ratio, that is, updating the concentration tilt in the sequence ALC to Ct=SRt×Ct0 / (1-SRt); and obtaining a risk-effect state value based on the phase autobias characteristic and the percentile value of the current measurement point in the normalized sequence ALC; Taking j1 as the serial number of the measuring point, in the liquid phase distance sequence ALP, the ratio of the deviation of any element to the element in the liquid phase distance sequence is recorded as the phase deviation value of the corresponding measuring point of the element; the phase deviation values ​​under each measuring point are normalized to form the deviation modulus sequence MdsL, and the phase self-bias PSd of the j1th measuring point can be calculated based on the liquid phase distance sequence and the deviation modulus sequence j1 , which is calculated as follows: PSd j1 =exp(ALP(j1)×MdsL(j1)); where exp() represents the exponential function with base e, ALP(j1) and MdsL(j1) represent the elements corresponding to the j1th measuring point in the liquid phase distance sequence and the offset modulus sequence, respectively.

2. The method for marking abnormal data of elemental nickel recovered from wastewater according to claim 1, characterized in that: In step S100, the method for initializing a recovery pool for recovering elemental nickel based on pulse current deposition technology is as follows: a transmission pipeline inputs the pretreated wastewater into the recovery pool, and a nickel ion concentration measuring device, a temperature sensor, and a pH sensor are arranged in the transmission pipeline to obtain concentration values, temperature values, and pH values ​​in real time.

3. The abnormal data marking method for recovering elemental nickel from wastewater according to claim 1, characterized in that: In step S300, the method of performing electronically controlled adaptation risk analysis to form a risk efficacy value based on the adaptation input feature set obtained in real time further includes: performing down-weighted fitting on the concentration tilt Ct0 that is 0.2 NIMD earlier than the current moment.

4. The method for marking abnormal data of elemental nickel recovered from wastewater according to claim 1, characterized in that: In step S400, the method for marking abnormal data for the collected data in combination with the risk effective state value is: within the current deposition monitoring interval, if the measuring point obtains the risk effective state value for the first time within the current deposition monitoring interval, the risk effective state value is recorded as the effective state base value of the measuring point, and the deposition setting group is obtained in real time, and the effective state base value and the deposition setting group are composed of a binary group as the adjustment identification group of the measuring point; wherein the deposition setting group is a multivariate array composed of actual measured values ​​corresponding to several deposition options, and the actual measured values ​​corresponding to the deposition options are obtained through a PID controller or a PLC controller, and the deposition options include: one or more of current, current density, voltage and electrode spacing; if a measuring point does not obtain the risk effective state value for the first time, the risk effective state value calculated for the measuring point is recorded as the effective state derivative value; among the various effective state derivative values ​​of any measuring point, the proportion of the effective state derivative values ​​with values ​​smaller than the effective state base value is taken as the effective state overflow ratio of the measuring point; A time period is set as the monitoring section TSS, with a value range of TSS∈[5, 10] minutes; the measuring point with an effective state overflow ratio exceeding 50% in the TSS period before the current moment is defined as the effective state reference measuring point at the current moment; the regulation identification group of the current measuring point and each effective state reference measuring point is subjected to box plot anomaly identification. If the current measuring point is an anomaly point, it will warn that there is a hysteresis risk in the performance of the current electrodeposition regulation system, and the measured values ​​of each deposition option obtained at that moment will be marked as abnormal data.

5. An abnormal data marking system for recovering elemental nickel from wastewater, characterized in that: The abnormal data marking system for recovering elemental nickel from wastewater includes: a processor, a memory, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, the steps of the abnormal data marking method for recovering elemental nickel from wastewater described in any one of claims 1-4 are implemented. The abnormal data marking system for recovering elemental nickel from wastewater runs on computing devices such as desktop computers, laptop computers, PDAs, and cloud data centers.

Images