Adaptive calibration method and system for soil heavy metal detection data

Abstract

The application relates to the technical field and discloses a self-adaptive calibration method and system for soil heavy metal detection data, which comprises the following steps: firstly, a pollution fingerprint database of a pretreatment device and a sample-device contact matrix are constructed, the element-specific total pollution contribution is calculated by combining the two, and a preliminary calibration value is obtained after deduction; secondly, a pollution interaction feature library is constructed through gradient concentration experiments and orthogonal experiments; finally, a device contact sequence is extracted, a dynamic interference quantity is calculated based on the feature library, and a final calibration value is obtained after deduction; the application can accurately calibrate detection data, effectively eliminates systematic deviation caused by trace accumulation of a pretreatment device, and avoids misjudgment of clean soil.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field, more particularly, the present application relates to a self-adaptive calibration method and system for soil heavy metal detection data. BACKGROUND

[0002] Soil heavy metal content detection is the core technical support for environmental quality assessment, land use planning and pollution control. The process standardization and equipment cleanliness of the pre-treatment link, as a key step before sample detection, directly affect the authenticity of the detection data. During the pre-treatment process, the sample needs to go through multiple processes such as grinding, concentration and purification. The long-term use of pre-treatment equipment is prone to produce trace pollution. This kind of pollution, after being accumulated through multiple processes, may significantly interfere with the detection results. Therefore, targeted calibration technology is needed to ensure data accuracy.

[0003] A soil heavy metal detection value correction method, device and computer storage medium are disclosed in Chinese patent with authorization announcement No. CN110018294B. By determining the factors affecting the detection value of the rapid detection instrument, the accurate value of laboratory analysis is taken as the benchmark to construct a mixed linear correction model. After parameter estimation and verification, the model is used to correct the detection value of the rapid detection instrument. The core purpose is to improve the detection accuracy of the rapid detection instrument itself. Chinese patent with authorization announcement No. CN101514980B proposes a method and device for rapidly detecting the content and spatial distribution of heavy metals in soil. The content is determined by a portable X-ray fluorescence spectrometer, the geographic coordinates are obtained by combining the global satellite positioning system, and the interference is deducted by calling the calibration coefficient corresponding to the soil type and water content. Finally, a visual spatial distribution map is generated, focusing on solving the problem of the influence of soil properties on the detection results.

[0004] However, the above-mentioned prior art does not pay attention to the problem of systematic deviation amplification caused by the "trace accumulation effect" of the pre-treatment equipment of the soil sample. The prior art CN110018294B focuses on the correction of the detection value of the detection instrument itself, and completely ignores the interference of the trace pollution caused by the wear and tear of the pre-treatment equipment, such as adsorption. The prior art CN101514980B only calibrates the sample properties such as soil type and water content, ignoring the pollution accumulation characteristics of multiple devices such as ball mill, rotary evaporator and ultrasonic cleaner in the pre-treatment link. Although the trace pollution of a single pre-treatment equipment (such as nickel and chromium ions released by the wear of grinding balls, and trace heavy metals adsorbed by glassware) is at a very low level, it can reach a detectable low concentration level after being accumulated through multiple processes. For low-concentration samples such as background soil, this accumulated pollution can significantly increase the detection results, leading to the misjudgment of clean soil as slightly polluted, and thus causing unnecessary soil remediation costs. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the present application provides a self-adaptive calibration method and system for soil heavy metal detection data, which quantifies equipment pollution contribution by constructing a pollution fingerprint database and a sample-equipment contact matrix, calculates a dynamic interference quantity by combining a pollution interaction feature library to realize hierarchical calibration; can effectively eliminate systematic deviation caused by trace accumulation of pretreatment equipment, significantly improve the detection accuracy of low-concentration samples, and avoid misjudgment of clean soil and unnecessary subsequent repair cost.

[0006] In order to achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] The self-adaptive calibration method for soil heavy metal detection data comprises:

[0008] A pollution fingerprint database of pretreatment equipment is constructed, a sample-equipment contact matrix is established, the original detection value of each heavy metal element in the soil sample to be calibrated is obtained, the element-specific total pollution contribution is calculated based on the pollution fingerprint database and the sample-equipment contact matrix, and the element-specific total pollution contribution is deducted from the original detection value of each heavy metal element in the soil sample to be calibrated to obtain the preliminary calibration value of each heavy metal element in the soil sample to be calibrated.

[0009] A pollution interaction curve is drawn through a gradient concentration experiment, an orthogonal experiment scheme is designed, and a sequence sensitivity index is calculated; a metal ion affinity matrix is constructed based on the pollution interaction curve and the sequence sensitivity index; the metal ion affinity matrix, the sequence sensitivity index and the pollution interaction curve are integrated to establish a pollution interaction feature library.

[0010] The equipment contact sequence of the soil sample to be calibrated is extracted from the sample-equipment contact matrix, the dynamic interference quantity of each heavy metal element is calculated based on the pollution interaction feature library and the equipment contact sequence of the soil sample to be calibrated, and the dynamic interference quantity is deducted from the preliminary calibration value to obtain the final calibration value of each heavy metal element in the soil sample to be calibrated.

[0011] The construction method of the pollution fingerprint database comprises: collecting pollutant samples at different spatial positions of a plurality of pretreatment equipment, detecting the pollutant samples to obtain a pollutant concentration data set of different pretreatment equipment, recording the sampling time, and integrating the pollutant concentration data sets of different pretreatment equipment to construct the pollution fingerprint database of the pretreatment equipment.

[0012] The method for establishing the sample-equipment contact matrix comprises:

[0013] A unique identification code is assigned to each soil sample. A position sensor is installed in the pretreatment equipment to establish a device position sensing network. When a soil sample with a unique identification code enters the pretreatment equipment, the device position sensing network records the pretreatment equipment number, the position coordinates of the soil sample in the pretreatment equipment, and the dwell time at each position coordinate.

[0014] Based on the residence time of the soil sample at each coordinate in the pretreatment equipment, the contact strength coefficient at each coordinate is calculated.

[0015] Using the unique identification code of the soil sample as the row and the position coordinates in the pretreatment equipment as the column, a sample-equipment contact matrix is ​​constructed. The contact intensity coefficient, residence time, and equipment number of each soil sample at different position coordinates in the pretreatment equipment are filled into the corresponding positions in the sample-equipment contact matrix.

[0016] The method for calculating the contact strength coefficient at each location coordinate includes:

[0017] If the residence time t of the soil sample at any coordinate in the pretreatment equipment is... stay The contamination migration threshold t of the corresponding pretreatment equipment is greater than thr Then, the contact intensity coefficient at the corresponding location coordinates is calculated based on the residence time and pollution migration threshold; if t stay Less than or equal to t thr , then I=0.

[0018] The calculation method for the element-specific total pollution contribution includes:

[0019] The contamination migration path of the soil sample to be calibrated can be obtained by querying the sample-equipment contact matrix based on the unique identification code of the soil sample to be calibrated.

[0020] For each contact point in the contamination migration path of the soil sample to be calibrated, the pollution contribution baseline value at the corresponding location is extracted from the contamination fingerprint database, and the pollution contribution baseline value is corrected for time to obtain the corrected pollution contribution value.

[0021] The total pollution contribution is obtained by traversing all contact points in the pollution migration path of the soil sample to be calibrated and summing the corrected pollution contribution values ​​of each contact point.

[0022] If the total pollution contribution is greater than 5% of the original detection value, then enter the deep calibration mode to obtain the element-specific total pollution contribution.

[0023] The method for time-correcting the pollution contribution baseline value includes:

[0024] judging an interval Δt of a contact time of any contact point in a pollution migration path of the soil sample to be calibrated and a sampling time, when the interval Δt is greater than a validity period t valid time-correcting a pollution contribution reference value of the corresponding contact point.

[0025] The method for drawing the pollution interaction curve comprises:

[0026] A series of gradient concentrations of the interaction source metal standard soil are prepared, and each concentration of the interaction source metal standard soil is divided into two groups for different equipment combination processing;

[0027] The heavy metal full spectrum analysis is performed on the interaction source metal standard soil processed by the two groups of different equipment combinations, the additional release amount of the interaction metal is calculated, and the pollution interaction curve is drawn with the interaction source metal concentration as the horizontal axis and the additional release amount of the interaction metal as the vertical axis.

[0028] The method for calculating the order sensitivity index comprises:

[0029] In the orthogonal test scheme, the use order of the pretreatment equipment is changed to generate a plurality of equipment combination schemes, the same standard soil sample is processed by each equipment combination scheme, the standard soil sample processed by different equipment combination schemes is detected, the detection concentration of each heavy metal is recorded, and an element-scheme-concentration three-dimensional data table is established.

[0030] The concentration set of each heavy metal element in all equipment combination schemes is extracted from the element-scheme-concentration three-dimensional data table, the highest concentration, the lowest concentration, and the average concentration of the current heavy metal element are calculated, and the order sensitivity index of the current heavy metal element is calculated according to the highest concentration, the lowest concentration, and the average concentration of the current heavy metal element.

[0031] The method for constructing the metal ion affinity matrix comprises:

[0032] The interaction influence intensity basic value is calculated based on the pollution interaction curve.

[0033] Whether there is metal interaction between the interaction source metal and the interaction metal is judged according to the pollution interaction curve.

[0034] The interaction source metal with the metal interaction is defined as the influencing party metal, and the interaction metal is defined as the influenced party metal, whether the influencing party metal and the influenced party metal are order sensitive elements is judged, if the influencing party metal and the influenced party metal are both order sensitive elements, the order correction is performed on the interaction influence intensity basic value according to the order sensitivity index of the influenced party metal to obtain the final interaction influence intensity value, otherwise, the interaction influence intensity basic value is directly taken as the final interaction influence intensity value.

[0035] According to the final interaction intensity value, a metal ion affinity matrix is constructed.

[0036] The adaptive calibration system of soil heavy metal detection data is used to implement the adaptive calibration method of soil heavy metal detection data, and the system comprises:

[0037] The pollution contribution calculation module is used to construct a pollution fingerprint database of a pretreatment device, establish a sample-device contact matrix, obtain the original detection value of each heavy metal element in the soil sample to be calibrated, calculate the element-specific total pollution contribution based on the pollution fingerprint database and the sample-device contact matrix, and calculate the element-specific total pollution contribution.

[0038] The preliminary calibration module is used to deduct the element-specific total pollution contribution from the original detection value of each heavy metal element in the soil sample to be calibrated to obtain the preliminary calibration value of each heavy metal element in the soil sample to be calibrated.

[0039] The pollution interaction quantification module is used to draw a pollution interaction curve through a gradient concentration experiment, design an orthogonal experiment scheme, calculate a sequence sensitivity index, construct a metal ion affinity matrix based on the pollution interaction curve and the sequence sensitivity index, integrate the metal ion affinity matrix, the sequence sensitivity index and the pollution interaction curve, and establish a pollution interaction feature library.

[0040] The final calibration module is used to extract the device contact sequence of the soil sample to be calibrated from the sample-device contact matrix, calculate the dynamic interference of each heavy metal element based on the pollution interaction feature library and the device contact sequence of the soil sample to be calibrated, and deduct the dynamic interference from the preliminary calibration value to obtain the final calibration value of each heavy metal element in the soil sample to be calibrated.

[0041] Compared with the prior art, the present application has the following beneficial effects:

[0042] The present application realizes the accurate quantification and stripping of the trace pollution accumulation of the pretreatment device by constructing the pollution fingerprint database and the sample-device contact matrix, breaks through the assumption limitation of the uniform pollution of the device in the traditional method, accurately captures the space-time non-uniformity characteristics of the device pollution, further integrates the nonlinear interference effects caused by the interaction between metals and the contact sequence of the device by means of the establishment of the pollution interaction feature library and the calculation of the dynamic interference, solves the problem of the amplification of the superposition of trace pollution under multi-process accumulation, and forms a complete closed loop from device pollution quantification to interaction interference calibration, significantly improves the accuracy of low-concentration soil sample detection, effectively eliminates the systematic deviation introduced by the pretreatment link, fundamentally avoids the misjudgment of clean soil as contaminated soil due to detection deviation, and reduces unnecessary subsequent repair costs. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description only aim to some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0044] Figure 1 A method flow chart of a soil heavy metal detection data adaptive calibration method provided by the embodiment of the present application;

[0045] Figure 2 A principle flow chart of constructing a sample-equipment contact matrix provided by the embodiment of the present application;

[0046] Figure 3 A principle flow chart of establishing a pollution interaction feature library provided by the embodiment of the present application;

[0047] Figure 4 A structure schematic diagram of a three-level pollution amplification model provided by the embodiment of the present application;

[0048] Figure 5 A function module diagram of a soil heavy metal detection data adaptive calibration system provided by the embodiment of the present application. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort belong to the protection scope of the present application.

[0050] Embodiment 1:

[0051] Please refer to Figure 1 The embodiment provides a soil heavy metal detection data adaptive calibration method, which comprises the following steps:

[0052] Step S10, constructing a pollution fingerprint database of a pretreatment equipment, establishing a sample-equipment contact matrix, obtaining original detection values of each heavy metal element in a soil sample to be calibrated, calculating element-specific total pollution contribution based on the pollution fingerprint database and the sample-equipment contact matrix, and deducting the element-specific total pollution contribution from the original detection values of each heavy metal element in the soil sample to be calibrated to obtain preliminary calibration values of each heavy metal element in the soil sample to be calibrated;

[0053] Further, step S10 comprises:

[0054] Step S11, collecting pollutant samples at different spatial positions of multiple pretreatment devices, detecting the pollutant samples, obtaining a pollutant concentration data set of different pretreatment devices, recording the sampling time, integrating the pollutant concentration data set of different pretreatment devices to construct a pollution fingerprint database of pretreatment devices; the pretreatment devices include a ball mill, a rotary evaporator and an ultrasonic cleaner;

[0055] Step S11 is to construct a pollution fingerprint database of pretreatment devices, which needs to collect pollutant samples at different spatial positions of multiple pretreatment devices, obtain a pollutant concentration data set and record the sampling time, and integrate the data to form a database. The pollutant sample refers to the heavy metal pollutants collected by a specific carrier at different spatial positions of the pretreatment device, which are generated by the wear, shedding or adsorption of the device itself. The selection of the carrier needs to meet the "no background interference" principle to ensure that the collected pollutants only come from the device itself. For the ball mill, high-purity blank quartz sand is selected as the sampling carrier, because quartz sand has stable chemical properties and does not contain target heavy metal elements, which can avoid interference of the carrier itself on the detection results. For the rotary evaporator, medical grade cotton swabs are selected as the sampling carrier, because the cotton swabs are soft and have high cleanliness, which can accurately wipe and collect trace residues on the surface of glassware. For the ultrasonic cleaner, blank filter paper is selected as the sampling carrier, because the filter paper has uniform adsorption performance and can quantitatively capture the trace metals suspended or shed in the cleaning tank.

[0056] The specific implementation process is that the inside of the ball mill is divided into multiple sampling points according to "upper-middle-lower part of the grinding tank", and the blank quartz sand samples after grinding are collected at a set time interval, and the first pollutant concentration data set is obtained through detection; the time interval is determined through pre-experiment: continuously monitor the pollution accumulation of blank quartz sand under different intervals, select the interval with stable pollution amount change and reflecting the real-time state of the equipment, and the exemplary interval is 1-4 hours. The evaporation flask of the rotary evaporator is divided into multiple regions according to "flask neck-flask body-flask bottom", and medical grade cotton swabs are used to wipe the surface of each region evenly, and the residues collected are dissolved and detected to obtain the second pollutant concentration data set; the cleaning tank of the ultrasonic cleaner is divided into multiple grid regions, and a blank filter paper sheet is placed in each grid region, and the filter paper sheet is taken out after the equipment runs for a set length of time, and the third pollutant concentration data set is obtained through digestion detection. The specific time point of each sampling is recorded synchronously as the sampling time. The first pollutant concentration data set covers multiple heavy metal elements to be detected, such as cadmium, mercury, chromium, nickel, lead, etc., and the specific concentration values at each sampling point of the ball mill; the second pollutant concentration data set covers the concentration values of the target heavy metal elements consistent with the first pollutant concentration data set in each region of the evaporation flask; the third pollutant concentration data set covers the concentration values of the target heavy metal elements consistent with the first pollutant concentration data set and the second pollutant concentration data set in each grid region, and the concentration data reflects the distribution of trace metals suspended or falling off in the cleaning tank. The three types of data sets are respectively bound and labeled with corresponding "equipment number-position coordinates-sampling time", and then all the labeled data are integrated to form a pollution fingerprint database.

[0057] In the prior art, the traditional calibration method relies on single-time-point and single-location pollution detection data, without considering the "spatiotemporal non-uniformity" of equipment pollution, resulting in insufficient data representation and inability to support accurate calibration. Step S11 solves the problem of "carrier background interference leading to distortion of pollutant concentration detection" through differential sampling carrier selection, ensuring that the collected data truly reflect the equipment pollution state; through internal multi-space location division and sampling, the problem of "single-location sampling cannot reflect the difference in internal pollution distribution of the equipment" is solved, so that the data can cover the spatial dimension characteristics of equipment pollution.

[0058] The targeted selection of the sampling carrier in step S11 controls the error of the contaminant concentration detection within a very low range, ensures the accuracy of the core data of the contamination fingerprint database, and captures the heterogeneity characteristics of the internal contamination of the equipment through multi-space position sampling. For example, the lower area of the ball mill releases more wear due to more intense collisions of the grinding balls than the upper area, which provides data support for the subsequent precise matching of the contamination reference value of the contact position of the sample. The introduction of the time tag enables the database to have a "time effectiveness screening" function, which can eliminate expired contamination data and avoid calibration deviation caused by changes in the contamination state of the equipment. The database formed by the integration of three-dimensional data realizes precise query of contamination data under the three-dimensional conditions of "equipment-position-time", and lays the foundation for extracting the contamination contribution reference value of the corresponding position in step S132. Step S11 breaks through the assumption limitation of the traditional mean deduction method for "uniform contamination" by constructing a database containing space-time information. Step S132 needs to extract the contamination contribution reference value of a specific space-time based on this database and make time corrections. Without step S11, step S132 will have no precise original data to call, and can only rely on experience to estimate the contamination contribution, resulting in significant deviation of the calibration result. Without step S11, the entire scheme will lack the basis for quantifying the characteristics of equipment contamination, and all subsequent calibration logic based on "spatial and temporal non-uniformity" cannot be established. The problem of random fluctuations in equipment contamination cannot be solved, and the risk of misjudgment of clean soil still exists.

[0059] Step S12, tracking the space-time trajectory of the soil sample in the pretreatment equipment, constructing a sample-equipment contact matrix;

[0060] Please refer to Figure 2 As shown in the figure, further, step S12 includes:

[0061] Step S121, assigning a unique identification code to the soil sample, installing a position sensor in the pretreatment equipment, and establishing an equipment position sensing network. When the soil sample with a unique identification code enters the pretreatment equipment, the pretreatment equipment number, the position coordinates of the soil sample in the pretreatment equipment, and the residence time at each position coordinate are recorded through the equipment position sensing network;

[0062] Step S122, based on the residence time of the soil sample at each position coordinate in the pretreatment equipment, calculating the contact intensity coefficient at each position coordinate;

[0063] If the residence time t stay of the soil sample at any position coordinate in the pretreatment equipment is greater than the contamination migration threshold t thr of the corresponding pretreatment equipment, the contact intensity coefficient at the corresponding position coordinate is calculated according to the residence time and the contamination migration threshold, and the specific formula is I=(t stay / t thr )×(1+R / Rstand ); otherwise, I = 0; wherein R is a real-time running parameter of the pre-treatment device, R stand is a standard running parameter of the pre-treatment device.

[0064] In step S123, a sample-device contact matrix is constructed with the unique identification code of the soil sample as the row and the position coordinates in the pre-treatment device as the column, and the contact intensity coefficient, residence time and device number of each soil sample at different position coordinates in the pre-treatment device are filled in the corresponding positions of the sample-device contact matrix.

[0065] Step S12 is to construct a sample-device contact matrix, which requires assigning a unique identification code to the soil sample, recording the position and residence time of the sample in the pre-treatment device through a position sensing network, calculating the contact intensity coefficient, and finally integrating the information in the form of a matrix. The specific implementation process is as follows: in step S121, the unique identification code adopts the "letter + number" combination coding rule, and the code needs to contain basic information such as sample batch and receiving time to ensure that the identification code of each sample is unique and traceable; position sensors are installed at key nodes of the pre-treatment device, and the type of the sensor is selected according to the characteristics of the device; an infrared position sensor is installed in the ball mill to monitor the position of the sample in the grinding tank; a laser position sensor is installed at the bottle opening of the rotary evaporator to track the area of the sample in the evaporating bottle; an array position sensor is installed in the ultrasonic cleaner to locate the grid position of the sample in the cleaning tank; all sensors are connected through an industrial bus to form a device position sensing network. When the soil sample enters the pre-treatment device, the device position sensing network captures the position coordinates of the soil sample in real time, and simultaneously records the residence time of the soil sample at each position coordinate through the device control system, i.e. the time difference from the entry of the soil sample into the position to the exit of the soil sample from the position.

[0066] In step S122, the calculation of the contact intensity coefficient I needs to determine the pollution migration threshold t thr firstly. thr The pollution migration threshold t thr is determined through a blank sample test: the blank sample is placed in different positions of the device, different residence times are set, and the heavy metal pollution of the sample is detected; when the increase of the pollution is greater than 3 times the detection error of the blank sample, it is determined that there is a "significant increase". For example, the ball mill has a faster pollution migration due to the grinding process, and the t thr value is 20-40 minutes; the rotary evaporator has a slower adsorption process, and the t thr value is 50-70 minutes; the ultrasonic cleaner has a faster migration due to vibration, and the t thr value is 10-20 minutes. The pollution migration threshold t thrThe essence is the minimum time threshold for effective migration of pollutants from the equipment to the sample, which is derived from the kinetic characteristics of physical and chemical processes such as adsorption, desorption, and diffusion. Pollutants on the surface of the equipment, such as metal ions from worn-out grinding balls and heavy metals adsorbed by glassware, need to overcome the interfacial energy barrier to migrate to the sample. Overcoming the energy barrier and subsequent migration accumulation all require time, so short contact time cannot complete effective migration and can only form negligible trace adhesion.

[0067] The real-time operating parameter is the actual operating parameter of the equipment during operation. The actual operating parameter of the ball mill is the real-time speed, the actual operating parameter of the rotary evaporator is the real-time temperature, and the actual operating parameter of the ultrasonic cleaner is the real-time power, which are all real-time retrieved from the equipment control system. The standard operating parameter is the optimal operating parameter set by the equipment manufacturer or the operating parameter with the lowest pollution release amount verified by multiple blank tests. When the residence time t stay of the sample in a certain position is greater than t thr , I = (t stay / t thr ) × (1 + R / R stand ); when t stay is less than or equal to t thr , I = 0. The formula logic is: t stay greater than t thr indicates that pollution migration has occurred, and the ratio of t stay to t thr reflects the degree of influence of residence time on pollution, and the greater the ratio, the higher the pollution risk. The ratio of R to R stand reflects the degree of deviation of the operating parameter from the standard state, and the greater the deviation, the more intense the wear or adsorption behavior of the equipment, and the higher the pollution release amount. The product of the two can comprehensively quantify the contribution strength of contact to pollution. The pollution migration threshold t thr is the minimum time threshold for effective migration of pollutants from the equipment to the sample, which is derived from the kinetic characteristics of physical and chemical processes such as adsorption, desorption, and diffusion. Pollutants on the surface of the equipment, such as metal ions from worn-out grinding balls and heavy metals adsorbed by glassware, need to overcome the interfacial energy barrier to migrate to the sample. Overcoming the energy barrier and subsequent migration accumulation all require time, so short contact time cannot complete effective migration and can only form negligible trace adhesion. Therefore, the relationship between t stay and t thr directly determines the presence and size of pollution contribution: t stay ≤ t thr , the pollutants have not completed effective migration, and the pollution amount is at a "negligible level"; t stay > t thr , the pollutants begin to migrate significantly, and the migration amount increases with the extension of the residence time. The core function of the contact intensity coefficient I is to quantify the "actual pollution contribution potential of the equipment to the sample". When t stay≤t thr When the pollution contribution potential is zero, directly assigning I = 0 can avoid complex micro-calculation, simplify the subsequent accumulation process of total pollution contribution, and avoid calibration deviation caused by "over-fitting of small signals". Through t thr Establishing clear boundaries between "effective pollution" and "ineffective pollution" can avoid misjudging negligible pollution caused by short-time contact as effective pollution, solve the problem of "over-deducting small pollution leading to low detection value" in traditional calibration, and improve the accuracy of low-concentration sample calibration. In step S123, the sample unique identification code is used as the row, and all preset position coordinates in the pretreatment equipment are used as the column to construct the sample-equipment contact matrix. The I value of the corresponding position, the residence time, and the equipment number are filled into the matrix by traversing the position record and contact intensity coefficient calculation result of each sample, and 0 is filled into the matrix for the positions not contacted, to form complete matrix data.

[0068] In the prior art, the traditional method relies on manual recording of sample processing path, which is prone to recording errors or omissions, and cannot quantify the contact degree of the sample and the equipment, resulting in lack of individual basis for pollution contribution calculation. Step S121 solves the problems of inaccurate association of sample and processing information and errors in manual recording of position and time by combining unique identification code and position sensing network, realizing automatic and accurate tracking of sample processing trajectory; step S122 solves the problem of ignoring the influence of contact time and equipment state on pollution by introducing t thr and running parameter correction, which makes the quantification of contact intensity more in line with the actual pollution migration law; step S123 solves the problem of scattered and chaotic contact information of multiple samples and multiple positions by integrating data in matrix form, realizing structured storage and batch calling of information. The unique identification code ensures that the processing trajectory of each sample can be traced individually, avoiding confusion of information between samples; the automatic recording of the position sensing network eliminates subjective errors of manual operation, and the recording accuracy of the position coordinates can reach centimeter level inside the equipment, and the recording error of the residence time can be controlled within seconds; the contact intensity coefficient is accurately distinguished from the pollution risk of different contact scenarios through double-factor correction, for example, the sample stays in the lower area of the ball mill for more than t thr and the speed is higher than R standWhen the I value increases significantly, it indicates that the contact point is a high pollution risk point; the matrix structure enables the subsequent steps to quickly locate the contact position and intensity data of the sample through the identification code, greatly improving the data query and calculation efficiency. The distribution of the contact intensity coefficient can be used to infer the high pollution risk area of the equipment, for example, if the I value of a specific position of the equipment is generally high, it may indicate that there is abnormal wear or pollution accumulation in that area, providing accurate guidance for local maintenance of the equipment; the matrix data can also be used to optimize the pretreatment process, by comparing the matrix data of different samples, the processing path with the lowest total contact intensity can be selected to reduce the cumulative risk of pollution from the source. Step S12 cooperates with step S31, and step S31 needs to extract the position of the non-zero I value from the matrix and sort it by time to construct the equipment transfer chain. If step S12 is missing, step S31 will not be able to obtain the equipment contact sequence of the sample, which will make the analysis of the pollution interaction effect lose its foundation. Without step S12, the processing trajectory and contact degree of the sample cannot be quantified, and the individual calibration of step S13 will degenerate into the traditional mean subtraction method, which cannot eliminate the pollution differences of individual samples; at the same time, the dynamic interference quantity calculation of step S30 cannot be carried out due to the lack of equipment contact sequence data, and the two core logics of the entire scheme, "spatial and temporal non-uniformity calibration" and "nonlinear interaction calibration", cannot be implemented.

[0069] In step S13, the original detection value of each heavy metal element in the soil sample to be calibrated is obtained, and based on the pollution fingerprint database and the sample-equipment contact matrix, the element-specific total pollution contribution is calculated. The element-specific total pollution contribution is subtracted from the original detection value of each heavy metal element in the soil sample to be calibrated to obtain the preliminary calibration value of each heavy metal element in the soil sample to be calibrated.

[0070] Further, step S13 includes:

[0071] In step S131, the original detection value of each heavy metal element in the soil sample to be calibrated is obtained, and the pollution migration path of the soil sample to be calibrated is obtained by querying the sample-equipment contact matrix according to the unique identification code of the soil sample to be calibrated.

[0072] The original detection value is obtained by inductively coupled plasma mass spectrometry or atomic fluorescence spectrometry. The choice of method depends on the characteristics of the target heavy metal element: inductively coupled plasma mass spectrometry is suitable for simultaneous detection of multiple elements with low detection limit, and is suitable for multiple trace heavy metals such as cadmium and mercury; atomic fluorescence spectrometry has higher sensitivity for elements such as arsenic and mercury, and can be selected according to detection needs. The unique identification code is a special identifier for the soil sample, which corresponds one-to-one with the row number of the sample-equipment contact matrix. By searching the identification code, the corresponding row in the matrix can be located, and the equipment number, position coordinates and residence time corresponding to all non-zero contact intensity coefficients in the row can be extracted. The trajectory formed by these information in time sequence is the pollution migration path.

[0073] Step S132, for each contact point in the pollution migration path of the soil sample to be calibrated, extracting the pollution contribution reference value of the corresponding position from the pollution fingerprint database, time correcting the pollution contribution reference value to obtain the corrected pollution contribution value;

[0074] The method of time correcting the pollution contribution reference value comprises:

[0075] Judging the interval Δt between the contact time of any contact point in the pollution migration path of the soil sample to be calibrated and the sampling time, when the interval Δt is greater than the effective period t valid , time correcting the pollution contribution reference value C base of the corresponding contact point.

[0076] The method of time correcting the pollution contribution reference value C base of the corresponding contact point comprises:

[0077] Constructing a time decay function containing a pollution decay time constant, and calculating the corrected pollution contribution value of the corresponding contact point according to the time decay function, the pollution contribution reference value of the corresponding contact point and the contact intensity coefficient of the corresponding contact point.

[0078] The pollution contribution reference value C base is the heavy metal concentration value at a specific space-time stored in the pollution fingerprint database, and the extraction process takes the "equipment number-position coordinates" of the contact point as the primary matching condition, locates all historical sampling records of the equipment at the position in the database, that is, multiple groups of "sampling time-C base " data pairs; and then takes the "contact time" of the sample at the position as the reference, calculates the interval Δt (Δt = |contact time-sampling time|) between the contact time and each sampling time, and selects the C base corresponding to the sampling time with the smallest Δt as the candidate value. The pollution index spectrum database accumulates multiple groups of C base of different sampling times under the same "equipment number-position coordinates" through continuous sampling at a set time interval, forms a C base sequence with "fixed position and dynamic time", and provides data basis for matching close to the actual contact time of the sample. The contact time refers to the specific time point at which the soil sample physically or chemically contacts at a certain position coordinate of the pretreatment equipment, and is the "process time data" generated in the soil sample pretreatment process, which is directly derived from the equipment position sensing network record of step S121.

[0079] The time correction needs to determine the effective period t valid , t valid is set through pre-experiment: for heavy metals with different stabilities, monitor their C baseThe amplitude of change over time, when the amplitude of change exceeds the detection error range, the corresponding time is the t of the element valid , for example, for volatile metals such as mercury, its C base decays quickly over time, t valid is short, with a value of 1-3 days; for stable metals such as lead, C base changes slowly, t valid is long, with a value of 5-10 days. When Δt is greater than t valid , it means that the candidate value has changed significantly due to the time being too long, and needs to be corrected. A time decay function f(Δt) = exp(-Δt / τ) is used for correction, where exp() is the exponential function with the base number e as the base number, and τ is the pollution decay time constant. The determination of τ is based on the volatility and stability of heavy metals: the τ of volatile metals is determined by monitoring its volatilization rate in the air, and the τ of stable metals is determined by monitoring its residual decay rate on the surface of the equipment. For example, for volatile metals, take a known concentration standard solution, uniformly coat it on the surface of the corresponding pretreatment equipment of typical material such as stainless steel ball mill and rotary evaporator glass, and place it in a temperature and humidity environment consistent with the actual pretreatment. Detect the surface residual concentration at a set time interval, and substitute the concentration change data over time into the exponential decay model to fit and obtain the time when the concentration decreases to 1 / e of the initial value, which is τ. For stable metals, the same standard solution is coated on the surface of the equipment material, simulating the standing / cleaning period after the equipment is used, and the residual amount is detected at regular intervals. Through decay model fitting, the time when the residual amount decays to 1 / e of the initial value is taken as τ.

[0080] The corrected pollution contribution value C corrected = C base × f(Δt) × I. C base is the basic pollution intensity of the coordinates of a certain position of the equipment, and is the calculation reference; f(Δt) quantifies the natural decay of C base with Δt, which conforms to the concentration change rule of volatile or stable metals; I reflects the driving effect of the actual contact between the sample and the equipment on the pollution migration by integrating the residence time and the equipment operation parameters. The multiplication of the three can completely restore the pollution contribution formation process of “reference intensity-time decay-contact driving”, which is logically self-consistent and conforms to the actual pollution mechanism. This formula precisely integrates the effects of spatial and temporal decay and contact strength, avoids the misestimation of pollution contribution caused by the expiration of C base or the difference in contact degree, makes the pollution contribution quantization more in line with the actual pollution of the sample, provides reliable data support for obtaining preliminary calibration values by deducting pollution, and effectively solves the defects of traditional calibration ignoring the differences in time and contact.

[0081] Step S133, all contact points in the pollution migration path of the soil sample to be calibrated are traversed, the corrected pollution contribution values of the contact points are accumulated, and a total pollution contribution is obtained; if the total pollution contribution is greater than five percent of the original detection value, a deep calibration mode is entered, and an element-specific total pollution contribution is obtained; otherwise, the total pollution contribution is directly taken as the element-specific total pollution contribution.

[0082] The deep calibration mode is: for strongly adsorbed elements, the total pollution contribution is multiplied by an enhancement coefficient a; for weakly adsorbed elements, the total pollution contribution is multiplied by a weakening coefficient b.

[0083] The total pollution contribution is the arithmetic sum of the pollution contributions of the contact points C corrected , reflecting the cumulative pollution amount of the sample after passing through all the contact points of the equipment. In the deep calibration mode, the strongly adsorbed elements refer to heavy metals that have a significantly higher adsorption amount on the surface of the pretreatment equipment than other elements due to physical and chemical properties, such as high ion charge number, ion radius matching the pore size of the equipment material, and chemical adsorption with the surface material of the equipment. The weakly adsorbed elements are the opposite. The distinction is determined by a blank adsorption experiment: a mixed solution of multiple heavy metals is contacted with a sample of the equipment material, and the adsorption amount of each element is detected. The top thirty percent of the adsorption amount is the strongly adsorbed element, and the bottom thirty percent is the weakly adsorbed element. The enhancement coefficient a is set according to the adsorption rate difference of the strongly adsorbed elements. The higher the adsorption rate, the larger the value of a. The weakening coefficient b is set according to the desorption rate of the weakly adsorbed elements. The faster the desorption rate, the smaller the value of b. The value range of a is 1.2 to 1.5, and the value range of b is 0.7 to 0.9. The element-specific total pollution contribution is obtained after correction.

[0084] Step S134, the element-specific total pollution contribution is deducted from the original detection value of each heavy metal element in the soil sample to be calibrated to obtain the preliminary calibration value of each heavy metal element in the soil sample to be calibrated.

[0085] In existing technologies, traditional calibration methods use a fixed mean to subtract contamination, failing to consider individual sample contamination pathway differences, the timeliness of contamination data, and element adsorption characteristics, leading to large calibration deviations. Step S131, through a combination of a unique identification code and matrix retrieval, solves the problem of "inaccurate matching between sample and contamination pathway," achieving personalized extraction of contamination pathways. Step S132, through a time correction mechanism, solves the problem of "calibration distortion caused by using expired contamination data," ensuring that contamination contribution calculation adapts to the time-varying nature of contamination on the equipment. Step S133, through a deep calibration mode, solves the problem of "misjudgment of contamination contribution due to ignoring differences in element adsorption," achieving element-differentiated calibration. Step S134, through accurate subtraction of contamination contribution, solves the problem of "original detection values ​​being unable to reflect the true content due to interference from equipment contamination." The adaptability of the detection method for the original detection values ​​ensures the accuracy of the initial data; the personalized extraction of contamination migration pathways ensures that the calibration of each sample is based on actual contact conditions, avoiding a "one-size-fits-all" mean error; the time correction mechanism filters out interference from expired data. corrected It can accurately reflect the actual pollution contribution at the current contact point; the deep calibration mode distinguishes element adsorption characteristics, making the element-specific total pollution contribution more consistent with the pollution accumulation patterns of different elements; the preliminary calibration value effectively removes the spatiotemporal non-uniformity of equipment pollution, significantly reducing the deviation from the true content of the sample. Step S13 works synergistically with steps S11 and S12 above: the pollution fingerprint database of step S11 provides C base The basic data, the sample-equipment contact matrix in step S12 provides the I value and the contamination migration path. The combination of these three elements forms a closed loop of "equipment contamination characteristics - sample contact degree - contamination contribution calculation". If the path extraction in S131 is missing, the subsequent contamination contribution calculation will not have a clear contact point. If the time correction in S132 is missing, outdated data will be used, resulting in calibration deviation. If the depth calibration in S133 is missing, it will not be able to adapt to the contamination deviation caused by elemental differences. If the subtraction calculation in S134 is missing, the calibration value of the contamination interference of the purification equipment cannot be obtained.

[0086] In step S10, the three-dimensional information structure of the contamination fingerprint database enables precise spatiotemporal characterization of equipment contamination, allowing for rapid retrieval of C values ​​at any given time and space. base The automated recording by the equipment position sensor network eliminates human error; the accuracy of position coordinate recording adapts to subtle differences in the internal contamination distribution of the equipment; and the error in residence time recording is controlled within a short time range. The contact intensity coefficient I, combining residence time and equipment operating parameters, accurately quantifies the potential contribution of contact to contamination. The preliminary calibration value effectively eliminates the spatiotemporal non-uniformity interference of equipment contamination, significantly reducing the detection deviation of low-concentration samples. (C from the contamination fingerprint database) base Time series changes can predict the degree of equipment wear and aging, for example, the C at a specific location of a certain piece of equipment. baseThe continuous increase indicates that there is abnormal wear at the position, and timely maintenance is required; the distribution of the I value of the sample-equipment contact matrix can optimize the pretreatment process, and by comparing the matrix data of different samples, the equipment contact path with the lowest I value sum is selected to reduce pollution accumulation from the source. Step S10 breaks through the assumption limitation of traditional calibration on "uniform pollution" by quantifying the degree of equipment pollution and sample contact; in cooperation with step S20, the preliminary calibration value of step S10 provides basic data for the nonlinear interactive calibration of S20. Without S10, S20 will lack initial data without interference of spatiotemporal non-uniformity, and cannot accurately analyze the "pollution resonance effect".

[0087] Step S20, draw the pollution interaction curve by gradient concentration experiment, design orthogonal experiment scheme, calculate the order sensitivity index; based on the pollution interaction curve and the order sensitivity index, construct the metal ion affinity matrix; integrate the metal ion affinity matrix, the order sensitivity index and the pollution interaction curve to establish the pollution interaction feature library;

[0088] Please refer to Figure 3 As shown in the figure, further, step S20 includes:

[0089] Step S21, prepare a series of gradient concentrations of interactive source metal standard soil, and divide each concentration of interactive source metal standard soil into two groups for different equipment combination processing;

[0090] The interactive source metal refers to the target heavy metal element set to explore its influence on other metals, the standard soil refers to the simulated soil matrix with known and uniform distribution of heavy metal element concentration prepared by artificial preparation, and the matrix selection needs to meet the "low background interference" principle, usually adopts high-purity quartz sand and clay mixed in proportion to ensure that the matrix itself does not contain target interactive source metal and interactive metal.

[0091] The specific implementation process is that the interactive source metal type is selected according to the common released metal type of the pretreatment equipment, such as chromium, nickel and the like, and a series of gradient concentration interactive source metal standard soil is prepared, the concentration gradient setting needs to meet the requirements of “covering the actual pollution release range and reflecting the dose effect”, and the concentration starting value and the gradient interval are determined through a pre-experiment: first, the actual concentration range of the interactive source metal released by the target pretreatment equipment is detected, the lower limit of the range is taken as the starting concentration C1, the adjacent concentration ratio is determined through a verification experiment, different ratios (such as 1.5, 2 and 3) are set for pre-experiments, the ratio that can clearly show the correlation between the concentration and the interaction effect is selected as the final interval, and the number of concentration gradients is required to be not less than 5, so as to ensure the accuracy of subsequent curve fitting. The standard soil at each concentration is divided into two groups, the first group adopts the “single equipment treatment” mode, and the core processing equipment in the target pretreatment equipment is selected, such as a ball mill, for treatment; the second group adopts the “multi-equipment combined treatment” mode, and the subsequent associated equipment is added on the basis of the first group of equipment, such as a ball mill + a rotary evaporator, and the parameters such as the treatment time and the operation parameters are kept consistent during the treatment of the two groups.

[0092] In the prior art, the traditional method does not focus on the influence of a single metal on other metals, and the source of the interaction cannot be distinguished under the interference of multiple metals, so that it is difficult to define the cause of the “pollution resonance effect”. The step S21 solves the problem that the dose-response relationship of the interaction effect cannot be constructed by means of “gradient concentration preparation”, so that the influence difference of different concentrations of the interactive source can be quantified; the problem that the multi-device superposition effect and the metal interaction effect cannot be separated is solved by means of the “two-group comparison treatment” technical means, and the influence of the metal interaction is separated from the device pollution superposition. The gradient concentration setting covers the actual pollution range of the equipment, and ensures the correlation between the experimental data and the actual scene; the selection of the single interactive source metal excludes the multi-source interference, so that the interaction effect detected subsequently can be clearly attributed to the target metal; the two-group parallel treatment controls the irrelevant variables, and ensures that the concentration difference is only caused by the metal interaction caused by the equipment combination.

[0093] In step S22, the interactive source metal standard soil treated by the two groups of different equipment combinations is subjected to heavy metal full spectrum analysis, the additional release amount of the interactive metal is calculated, and a pollution interaction curve is drawn with the concentration of the interactive source metal as the horizontal axis and the additional release amount of the interactive metal as the vertical axis; whether there is metal interaction is judged according to the pollution interaction curve.

[0094] The metal affected by the interaction source metal is a heavy metal element, and the additional release amount ΔNI is the difference between the concentration of the metal affected by the interaction in the multi-device combination processing group and the single device processing group, which is used to quantify the additional effect caused by the interaction source metal.

[0095] The specific implementation process is that all gradient concentration samples after processing are subjected to heavy metal full spectrum analysis. Inductively coupled plasma mass spectrometry is used because it can simultaneously detect multiple metal elements and has low detection limit, which is suitable for the detection requirements of trace metal interaction effect. After detection, the data is recorded according to the "concentration gradient-processing group-metal type" classification. For each concentration gradient, the concentration value of the metal affected by the interaction is extracted: the first group concentration is denoted as C single , the second group concentration is denoted as C multi , and the additional release amount ΔNI=C multi -C single . The interaction source metal concentration is taken as the abscissa, and ΔNI is taken as the ordinate. The pollution interaction curve is drawn by using curve fitting methods such as polynomial fitting and nonlinear regression. The fitting process needs to ensure that the goodness of fit R 2 meets the set requirements. Whether there is interaction between metals needs to meet two conditions: one is that ΔNI is significantly greater than 0. The standard deviation of the detection value is calculated by repeatedly detecting the blank sample to determine the random error threshold, or the detection limit of the detection method is directly used as a reference benchmark. When ΔNI is greater than 3 times the random error threshold, or ΔNI is greater than 2 times the detection limit, it is determined that ΔNI is significantly greater than 0, and accidental fluctuation interference is excluded. The second is that ΔNI and the interaction source metal concentration have a dose-dependent relationship: if the pollution interaction curve shows a non-horizontal trend, that is, with the increase of the interaction source metal concentration, ΔNI shows an explicit and repeatable change rule of increasing, decreasing or first increasing and then decreasing, rather than random fluctuations, it is proved that the interaction effect has dose-dependent relationship.

[0096] In the prior art, the traditional method cannot quantify the interaction intensity between metals, and cannot verify the existence of the "pollution resonance effect", resulting in unclear mechanism of multi-metal cross interference. Step S22 solves the problem of inaccurate detection of the concentration of the interacted metal through heavy metal full spectrum analysis, realizes the simultaneous quantification of multiple elements; solves the problem of separation of interaction effect and basic pollution through additional release amount calculation, accurately extracts the concentration change caused by metal interaction; solves the problem of verification of the existence of interaction effect through curve drawing and double condition judgment, eliminates random errors and confirms the dose dependence. Full spectrum analysis realizes the simultaneous detection of multiple interacted metals, and can explore the influence of one interaction source on multiple metals at one time; the calculation of ΔNI eliminates the interference of basic pollution through comparison of two groups, and only retains the additional effect caused by interaction; double condition judgment ensures the authenticity of the interaction effect, avoiding the misjudgment of detection error as interaction. The "inflection point" of the pollution interaction curve can reveal the nonlinear characteristics of the interaction effect, most curves show slow growth at low concentration and rapid growth at high concentration, and the inflection point, which explains why trace accumulation effect can cause systematic deviation amplification, that is, when the interaction source metal accumulates to the inflection point concentration, the release amount of the interacted metal increases nonlinearly; at the same time, the curve shape can distinguish the interaction type (promotion or inhibition), if ΔNI increases with the increase of concentration, it is a promotion effect, and if ΔNI decreases with the increase of concentration, it is an inhibition effect, which provides a basis for subsequent differentiated calibration. Step S22 is directly coordinated with the previous step S21, and the two groups of gradient samples provided by S21 are the basis for the detection and calculation of S22, and the curve results of S22 verify whether the interaction source metal selected by S21 exists in reality; and the subsequent steps S23 and S24 form a link, after S22 verifies the existence of interaction, the sequential experiment of S23 and the construction of affinity matrix of S24 have significance. If step S22 is missing, it cannot be confirmed whether there is interaction between metals, and the experiments of S23 and S24 will lose the logical premise, which may lead to waste of invalid experimental cost, and the core problem of "pollution resonance effect" cannot be verified, and the subsequent calibration scheme lacks pertinence.

[0097] Step S23, design an orthogonal experiment scheme, change the use order of the pretreatment equipment, generate multiple equipment combination schemes, process the same standard soil sample with each equipment combination scheme, and calculate the sequential sensitivity index;

[0098] The method for calculating the order sensitivity index comprises: detecting standard soil samples processed by different device combination schemes, recording the detection concentration of each heavy metal, and establishing an element-scheme-concentration three-dimensional data table; extracting the concentration set of each heavy metal element in all device combination schemes from the element-scheme-concentration three-dimensional data table, calculating the highest concentration, the lowest concentration and the average concentration of the current heavy metal element; calculating the order sensitivity index of the current heavy metal element according to the highest concentration, the lowest concentration and the average concentration of the current heavy metal element; if the order sensitivity index of the current heavy metal element is greater than a preset sensitivity threshold, the current heavy metal element is determined to be an order sensitive element, otherwise, the current heavy metal element is determined to be a non-order sensitive element; traversing all heavy metal elements recorded in the element-scheme-concentration three-dimensional data table, calculating the order sensitivity index of each heavy metal element one by one, and finally outputting the list of order sensitive elements and the list of non-order sensitive elements.

[0099] Specifically, the orthogonal test scheme refers to a scheme that covers all device order combinations with the least number of experiments by reasonably designing the experiments, so as to ensure that the influence of the device order can be fully explored; the order sensitivity index S refers to a dimensionless parameter representing the fluctuation degree of the concentration of the heavy metal element with the change of the device order, reflecting the response characteristics of the heavy metal element to the device order; the order sensitive element refers to an element whose S is greater than a preset sensitivity threshold, and is the main participant of the “pollution resonance effect”.

[0100] The specific implementation process is as follows: according to the core devices in the actual pretreatment process, the types of pretreatment devices participating in the experiment are determined, such as a ball mill, a rotary evaporator and an ultrasonic cleaner; then an orthogonal experiment scheme is designed to generate multiple device combination schemes, which need to cover all possible device order arrangements to ensure that each order has a corresponding experimental group, and the number of groups is determined according to the number of devices, for example, 6 order combinations correspond to 3 devices, 24 order combinations correspond to 4 devices, and exemplary device combination schemes include device combination scheme 1 (ball mill-ultrasonic-evaporation), device combination scheme 2 (ultrasonic-ball mill-evaporation), device combination scheme 3 (ball mill-evaporation-ultrasonic), device combination scheme 4 (evaporation-ball mill-ultrasonic), etc. The standard soil sample uses the same low background matrix as S21, and a plurality of target heavy metal elements with known concentrations are added in advance, and the addition concentration is referenced from the soil background value and the device pollution accumulation range to ensure that the detection concentration is within the linear interval of the method. The standard soil sample is processed by each device combination scheme, and the irrelevant variables such as device operating parameters and processing time are kept consistent during the processing; after the processing is completed, all samples are detected for heavy metals, and the inductively coupled plasma mass spectrometry method is used to record the data according to the dimensions of “heavy metal element-device combination scheme-detection concentration”, and an element-scheme-concentration three-dimensional data table is established.

[0101] For each heavy metal element, the sequence sensitivity index is calculated separately: from the element-treatment-concentration three-dimensional data table, the concentration set of the heavy metal element in all equipment combination treatments is extracted, the highest concentration C max, the lowest concentration C min and the average concentration C average are determined, C max is the maximum value in the concentration set, C min is the minimum value in the concentration set, and C average is the arithmetic mean of the concentration set, C max, C min and C average are substituted into the formula S = (C max-C min) / C average, and the sequence sensitivity index S of the element is calculated. In the formula, the numerator is the fluctuation range of the element concentration, the denominator is the average level of the element concentration, the ratio reflects the proportion of the concentration fluctuation amplitude relative to the average level, and the dimensionless. The core requirement of step S23 is to judge whether the concentration of the heavy metal element fluctuates significantly with the change of the pretreatment equipment sequence, i.e., whether it is sequence sensitive or not. The traditional method relies on subjective judgment and lacks objective indicators, while the formula converts the abstract "sensitivity degree" into a quantifiable dimensionless index through "the ratio of extreme difference to average value", solving the problem of "no unified standard for sensitivity judgment". The influence of equipment sequence change on element concentration is reflected not only in the fluctuation amplitude, but also in the proportion of fluctuation relative to its own concentration. In low-concentration background soil, even a small absolute value fluctuation (such as 0.01 mg / kg) may cause misjudgment if the proportion of the average concentration is high, such as 50%; while the same absolute value fluctuation of high-concentration elements has a low proportion, and the influence can be ignored. The numerator (C max-C min) in the formula quantifies the "fluctuation amplitude", and the denominator (C average) normalizes it to "relative proportion", which covers both absolute and relative values, and accurately matches the needs of different concentration levels of elements to distinguish the sensitivity difference of sequence. The setting of the sensitivity threshold is determined by pre-experiment: detect the elements with no sequence sensitivity, calculate their S values, and take 1.5 times of the value as the sensitivity threshold.

[0102] In the prior art, the traditional method defaults that the device sequence has no effect on pollution, and does not consider the nonlinear amplification caused by the change of the sequence, resulting in inconsistent calibration deviations of samples with different processing sequences. Step S23 solves the problem that the influence of the device sequence cannot be fully explored through orthogonal experimental design, ensuring that all sequence combinations are covered; solves the problem of chaotic data of multiple samples, multiple elements and multiple schemes through three-dimensional data table construction, realizing structured management of data; solves the problem that the sensitivity of elements to the sequence cannot be quantified and distinguished through sequence sensitivity index calculation and threshold determination, and clarifies the key participants of the resonance effect. The orthogonal experiment realizes full sequence coverage with the least number of times, reduces the experimental cost and ensures the integrity of the data; the three-dimensional data table clearly presents the changes of element concentration with the scheme, which is convenient for subsequent analysis; the sequence sensitivity index quantifies the sensitivity, avoiding subjective judgment errors; the threshold determination realizes accurate screening of sensitive elements. Step S23 cooperates with the foregoing step S22, S22 verifies the interaction between the metals, providing a prerequisite for S23 to explore the influence of the device sequence on the interaction; cooperates with the following step S24, the sequence sensitivity index and the list of sensitive elements of S23 provide core parameters for the affinity matrix modification of S24. If step S23 is missing, the influence of the device sequence on different elements cannot be clarified, the affinity matrix of S24 cannot be modified in the sequence dimension, resulting in that the subsequent calibration cannot adapt to the nonlinear interference caused by the process change; at the same time, the main participants of the resonance effect cannot be identified, and the calibration will fall into the dilemma of “comprehensive modification but unclear focus”, which cannot efficiently eliminate the core interference.

[0103] In step S24, based on the pollution interaction curve and the sequence sensitivity index, an affinity matrix of metal ions is constructed.

[0104] Further, step S24 comprises:

[0105] In step S241, based on the pollution interaction curve, an interaction influence intensity basic value is calculated.

[0106] In the pollution interaction curve, n1 feature concentration points are selected, the unit influence amplitude of each feature concentration point is calculated, the mean value of the unit influence amplitude of each feature concentration point is calculated as the average influence amplitude, and the average influence amplitude is normalized to obtain the interaction influence intensity basic value.

[0107] The method for normalizing the average influence amplitude is as follows: a blank control group without an interaction source metal is designed, and blank control group data is obtained, that is, the natural release amount of the standard soil containing only the interacted metal after being processed by the same device, and the interaction influence intensity basic value is calculated according to the natural release amount, that is, the average influence amplitude divided by the natural release amount.

[0108] Step S242, define the interaction source metal existing intermetallic interaction as the influencing party metal and the interacted metal as the influenced party metal, if both the influencing party metal and the influenced party metal are sequence-sensitive elements, then according to the sequence-sensitive index of the influenced party metal, sequence-correct the interaction influence intensity basic value to obtain the final interaction influence intensity value; otherwise, directly take the interaction influence intensity basic value as the final interaction influence intensity value, and according to the final interaction influence intensity value, construct the metal ion affinity matrix A, the matrix element A ij represents the interaction influence intensity value of the influencing party metal i on the influenced party metal j.

[0109] Specifically, the characteristic concentration point refers to a concentration node that can reflect the key change law of the pollution interaction curve, such as the minimum concentration point, the inflection point, and the maximum concentration point. The inflection point needs to be determined by calculating the second-order derivative of the curve, that is, the concentration point at which the second-order derivative is zero, which can accurately capture the turning feature of the interaction effect from slow change to rapid change. The unit influence amplitude refers to the ratio of the additional release amount of the interacted metal to the concentration of the interaction source metal at a single characteristic concentration point, reflecting the interaction intensity caused by a unit concentration of the interaction source. The average influence amplitude is the arithmetic mean of the unit influence amplitudes of all characteristic concentration points, eliminating the accidental error of a single concentration point. The interaction influence intensity basic value is the average influence amplitude normalized by the blank control group, used to eliminate the interference of the natural release of the interacted metal. The sequence correction coefficient is a parameter for adjusting the basic value in combination with the sequence-sensitive index, reflecting the amplification or weakening effect of the device sequence on the interaction effect of sensitive elements. The metal ion affinity matrix is a two-dimensional matrix with the influencing party metal as the row and the influenced party metal as the column, and the matrix element represents the interaction intensity of a specific metal pair.

[0110] The specific implementation process is as follows: in step S241, first select n1 characteristic concentration points from the pollution interaction curve. The determination of n1 is based on the complexity of the curve, and the minimum number that can make the average influence amplitude error less than the detection error is selected through pre-experiment verification for different forms of curves, such as linear, nonlinear, and S-shaped. For example, 3-5 characteristic points are selected, covering the minimum concentration, all inflection points, and the maximum concentration. Calculate the unit influence amplitude of each characteristic concentration point = additional release amount of the interacted metal ÷ concentration of the interaction source metal, and then calculate the arithmetic mean of these unit influence amplitudes as the average influence amplitude. Design a blank control group without the interaction source metal, and the blank control group is a standard soil containing only the interacted metal, which is treated by the same device combination as the experimental group. Detect the natural release amount of the interacted metal, divide the average influence amplitude by the natural release amount to obtain the interaction influence intensity basic value.

[0111] In step S242, the interaction source metal initiating the interaction is defined as the influencing party metal, and the metal affected by the interaction is defined as the affected party metal. It is determined whether the influencing party and the affected party are both sequence-sensitive elements, as indicated in the interaction intensity correction condition based on element sensitivity in Table 1. If both are sequence-sensitive elements, a sequence correction coefficient is set based on the sequence sensitivity index of the affected party. The sequence correction coefficient is set according to the S value of the affected party: a corresponding relationship between the S value of the affected party and the sequence correction coefficient is established through a pre-experiment. The larger the S value of the affected party, the larger the sequence correction coefficient. For example, when the S value of the affected party is greater than 0.3, the correction coefficient is 1.2-1.5, and when the S value of the affected party is between 0.2 and 0.3, the correction coefficient is 1.0-1.2. If there is a non-sensitive element, no correction is performed, and the basic value of the interaction intensity is directly used as the final interaction intensity value. The final interaction intensity value is calculated according to the sequence correction coefficient: final interaction intensity value = interaction intensity basic value x sequence correction coefficient. The influencing party metal is used as the row index, and the affected party metal is used as the column index to construct a metal ion affinity matrix A, and the final interaction intensity value is filled into the corresponding position of the matrix A ij , A ij represents the interaction intensity of the influencing party metal i on the affected party metal j, where i is the row index of the metal ion affinity matrix, and j is the column index of the metal ion affinity matrix.

[0112] Table 1 Interaction intensity correction condition based on element sensitivity

[0113]

[0114] When both are sensitive elements, the resonance condition of "sequence sensitivity + interaction" is formed: the release amount of the influencing party fluctuates with the sequence, the concentration of the affected party fluctuates with the sequence, and the interaction becomes a bridge for the fluctuation transmission. The correction coefficient is set by selecting the S value of the affected party because the affected party is the final bearer of the interaction effect, and the interaction effect is ultimately reflected in the concentration change of the affected party. The S value of the affected party reflects the receiving and amplifying ability of the interaction effect. Only when the S value of the affected party is large, can the fluctuation of the influencing party be converted into a significant change in the concentration of the affected party through the interaction. The S value of the influencing party only represents the sequence stability of its own release, and is irrelevant to the response of the affected party to the interaction. The sequence correction coefficient needs to quantify the amplification effect of the sequence on the interaction effect, so the S value of the affected party is used as the basis to accurately capture the key variables of resonance and avoid confusion between source fluctuation and effect amplification ability.

[0115] The essence of the sequence sensitivity index S is the fluctuation degree of the heavy metal element concentration with the change of the pretreatment equipment sequence. The greater the S value, the more intense the response of the element to the equipment sequence, that is, a slight adjustment of the equipment sequence will cause a significant fluctuation of the concentration of the element. The superposition of this fluctuation characteristic and the intermetallic interaction effect is the core cause of the "pollution resonance effect": the influencing party metal changes the adsorption / release behavior of the influenced party metal through interaction, and the high sequence sensitivity of the influenced party metal amplifies the final manifestation of this interaction effect. When the influencing party metal triggers interaction, the result of its action needs to be presented through the concentration change of the influenced party metal. If the S value of the influenced party is large, it means that the equipment sequence has become a key variable for regulating its concentration. The influencing party metal released by the front equipment will interact with the influenced party metal through the material characteristics of the rear equipment, such as the adsorption of positively charged ions by glass, and the high sensitivity of the influenced party will superimpose and amplify the "interaction" and "sequence fluctuation" effects. At this time, setting a larger sequence correction coefficient is to quantify this "interaction + sequence sensitivity" resonance amplification effect, so that the final interaction influence intensity value A ij is consistent with the actual pollution law. Conversely, if the S value of the influenced party is small, its concentration is weakly affected by the sequence, and even if there is interaction, the amplification effect of the sequence on the interaction result can be ignored, so the sequence correction coefficient increases with the increase of the S value, which is a precise adaptation to the law that "the resonance effect intensity is positively correlated with the sensitivity of the influenced party".

[0116] Non-sensitive elements refer to elements with an S value less than or equal to the sensitivity threshold, and the core feature is that the concentration is not affected by the equipment sequence. When either the influencing party or the influenced party is a non-sensitive element, the amplification of the sequence on the interaction effect loses its basis. When the influenced party is a non-sensitive element, its concentration has no response to the sequence, and even if the influencing party is a sequence-sensitive element, the interaction effect only manifests as a change in the basic adsorption / release, without additional fluctuations caused by the sequence. Forced correction will introduce a false amplification effect, leading to distortion of the interaction influence intensity value. When the influencing party is a non-sensitive element, its release amount is minimally affected by the sequence, and the interaction source intensity is stable, even if the influenced party is a sequence-sensitive element, its concentration fluctuation is only due to its own response to the sequence, and has nothing to do with the interaction. Correction will confuse its own fluctuation and resonance fluctuation, causing calibration bias. Therefore, when one party is a non-sensitive element, the interaction intensity is determined only by the inherent characteristics of the metal, and there is no need for correction. Directly using the basic value of the interaction influence intensity can avoid excessive correction errors.

[0117] In the prior art, the traditional method cannot quantify the interaction intensity between metals, and does not consider the influence of device sequence on interaction, resulting in that the multi-metal cross interference cannot be accurately calibrated. Step S241 solves the problems of single concentration point interference and natural release confusion in interaction intensity calculation by selecting and normalizing the feature concentration point, so that the basic value only reflects the additional interaction caused by the interaction source; step S242 solves the problem that the asymmetry and sequence sensitivity of metal interaction cannot be integrated into the quantitative system by sequence correction and matrix construction, realizing the structured presentation of the interaction law. The targeted selection of the feature concentration point captures the key information of the curve, and the average influence amplitude reduces the accidental error; the normalization processing eliminates the natural release interference, and the basic value can accurately reflect the essence of interaction; the sequence correction coefficient combines the sensitive characteristics of the affected party, so that the final interaction influence intensity value adapts to the actual influence of the device sequence; the matrix structure clearly presents the interaction relationship between the metal pairs, which is convenient for quick query and call.

[0118] Step S25 integrates the metal ion affinity matrix, sequence sensitivity index and pollution interaction curve to establish a pollution interaction feature library.

[0119] Step S25 is to integrate the metal ion affinity matrix, sequence sensitivity index and pollution interaction curve to establish a pollution interaction feature library. The pollution interaction feature library refers to a structured database systematically storing the interaction law between metals, element sequence sensitivity and device influence characteristics, having data query, update and reuse functions, and its core value is to convert scattered experimental data and quantitative indexes into an engineering tool that can directly serve calibration.

[0120] The specific implementation process is to first determine the core data architecture of the feature library, and design the storage structure according to the three-dimensional dimensions of "metal pair-device sequence-interaction feature": the first dimension is "affected metal-affected metal" pair, which is associated with A ijThe second dimension is "device combination scheme", and the corresponding sequence-sensitive index S and the shape parameters of the pollution interaction curve, such as the inflection point concentration and the maximum slope, are associated. The third dimension is "device inherent characteristics", and the main metal element set released by the front device and the metal element set easily adsorbed by the rear device are associated. These sets are derived from the pollution fingerprint database in step S11, and are obtained by screening the top 3 metal release or adsorption concentrations of each device. The metal ion affinity matrix in step S24 is disassembled into entries according to the metal pairs, and is bound with the pollution interaction curve of the corresponding metal pair and the sequence-sensitive index of each metal, and then is associated with the device combination scheme and the device inherent characteristics to form a complete data entry. A data indexing mechanism is established to support multi-condition queries such as "metal pair + device combination" and "sensitive element + device type". Step S25 solves the problem of systematic integration of multi-dimensional interaction data through three-dimensional architecture design, realizes the ordered storage of data, and clearly associates metal interaction, sequence influence and device characteristics to avoid data fragmentation.

[0121] The traditional calibration uses a linear superposition method to calculate the pollution contribution, without identifying the nonlinear interaction between metals, and ignoring the amplification effect of device sequence on the interaction, resulting in significant deviations in the detection results of low-concentration samples due to resonance effect. Step S20 constructs a pollution interaction feature library through gradient experiments and orthogonal experiments, which mainly solves the "pollution resonance effect" between the released metals of different pretreatment devices, that is, the nonlinear interaction between metals and the amplification effect of device sequence on the interaction. Gradient experiments and comparative processing ensure the pertinence of interaction data, curve judgment ensures the rigor of interaction verification, orthogonal experiments ensure the comprehensiveness of sequence influence, matrix construction ensures the accuracy of interaction quantification, and feature library integration ensures the efficiency of rule reuse. Step S20 and step S30 form a link, and the pollution interaction feature library in step S20 provides core parameters for step S30 to calculate the interaction enhancement factor and the dynamic interference quantity. Without step S20, the pollution resonance effect cannot be identified and quantified, and the calibration in step S30 can only eliminate linear cumulative pollution and cannot handle nonlinear amplified interference, so clean soil may be misjudged as contaminated soil due to resonance effect.

[0122] In step S30, the device contact sequence of the soil sample to be calibrated is extracted from the sample-device contact matrix, and based on the pollution interaction feature library and the device contact sequence of the soil sample to be calibrated, the dynamic interference quantity of each heavy metal element is calculated, and the final calibration value of each heavy metal element in the soil sample to be calibrated is obtained by deducting the dynamic interference quantity from the preliminary calibration value.

[0123] Further, step S30 includes:

[0124] In step S31, the device contact sequence of the soil sample to be calibrated is extracted from the sample-device contact matrix, and the continuous device pairs are identified to construct the device transfer chain.

[0125] Locate the positions corresponding to all non-zero contact strength coefficients in the sample-equipment contact matrix, which represent the actual contact of the soil sample with the specific region of the equipment; then retrieve the timestamps recorded by the position sensing network, sort these positions in chronological order to form the equipment contact sequence; then traverse the equipment contact sequence to identify adjacent equipment combinations as consecutive equipment pairs, such as a ball mill followed by a rotary evaporator, and record the difference between the end time of the former equipment and the start time of the latter equipment in the consecutive equipment pair as the transfer time interval; finally, connect all consecutive equipment pairs in order, label the type, contact position coordinates, and transfer time interval of each equipment, and complete the construction of the equipment transfer chain.

[0126] The prior art does not pay attention to the pollution interaction in the equipment transfer process, and defaults that the pollution of each equipment is independent, which leads to the fact that the pollution resonance effect cannot be captured. Step S31 solves the problem of sample processing trajectory fragmentation and the inability to restore the complete interaction scenario by extracting the equipment contact sequence, which makes the relationship between equipment clear; by constructing the equipment transfer chain, the transfer process of the consecutive equipment pair, which is the carrier of the pollution resonance effect, is clarified, providing a concrete analysis object for subsequent quantification of the interaction amplification effect. Step S31 cooperates with step S12: the sample-equipment contact matrix of step S12 provides a data basis for extracting non-zero contact positions, while the sequence sorting and chain construction of step S31 activates the value of the time dimension in the matrix, combining static contact strength data with dynamic processing flow. If this step is missing, the subsequent steps will not be able to locate the specific equipment combination where the interaction occurs, and the analysis of the "pollution resonance effect" will lose its targeting, resulting in the inability to carry out steps S32 to S34 due to the lack of a clear analysis object.

[0127] Step S32, based on the equipment transfer chain and the pollution interaction feature library, calculates the interaction enhancement factor of each consecutive equipment pair;

[0128] For each consecutive equipment pair in the equipment transfer chain, first retrieve the set of main metal elements released by the former equipment and the set of metal elements easily adsorbed by the latter equipment from the pollution interaction feature library. The set of main metal elements released by the former equipment is derived from the top 3 elements with the highest average release concentration of each metal in the pollution fingerprint database of step S11, and the set of metal elements easily adsorbed by the latter equipment is derived from the top 3 elements with the highest adsorption concentration of each region in the pollution fingerprint database; then calculate the intersection of the two sets, and the intersection elements constitute the metal element pair involved in the interaction; for the kth metal element pair involved in the interaction ( , ), extract the interaction influence strength value , of the kth metal element pair from the metal ion affinity matrix of the pollution interaction feature library ; and Substitute the interaction enhancement factor calculation formula:

[0129] ;

[0130] Get the interaction enhancement factor, where:

[0131] : the interaction enhancement factor of the continuous device pair, representing the amplification degree of the resonance effect in the device pair, ≥1, γ =1 indicates no additional interaction amplification, γ >1 indicates that there is amplification;

[0132] n: the total number of metal element pairs participating in the interaction in the current continuous device pair, calculate the intersection of the main release set of the front device and the easy adsorption set of the back device, the number of metal pairs formed by the intersection elements;

[0133] k: the index of the metal element pair participating in the interaction in the current continuous device pair;

[0134] : the influencing party metal in the kth group of metal element pairs participating in the interaction;

[0135] : the influenced party metal in the kth group of metal element pairs participating in the interaction;

[0136] : the transfer time interval of the continuous device pair;

[0137] : the order sensitivity index of the influenced party metal in the kth group of metal element pairs participating in the interaction;

[0138] : the interaction decay time constant, representing the characteristic time of metal ion decay with transfer time, determined by pre-experiment: monitor the change of metal interaction effect under different transfer time intervals, select the characteristic time corresponding to the stable decay rate, the exemplary value is 2 hours.

[0139] The interaction enhancement factor is the core parameter for quantifying the three-dimensional coupling effect of "intrinsic metal interaction + time decay + order sensitivity" in a single continuous device pair. Its core role is to convert the abstract "pollution resonance effect" into a calculable engineering index, which specifically represents: during the transfer of metal ions released by the front device to the subsequent device, the combined effect of metal affinity, ion decay caused by transfer time, and the sensitivity of the influenced metal to device order, causing the amplification of the influenced metal pollution contribution.

[0140] , it means that the influencing party metal has no additional interaction effect on the influenced party metal, at this time​ , the formula result , corresponding to the baseline state without interaction amplification, consistent with the objective law that there is no amplification without interaction. The baseline value is used to peel off the non-interaction, only the additional interaction effect is retained, that is, the "incremental amplification" part caused by interaction, to avoid confusing the "basic pollution contribution" with the "interaction amplification contribution", solving the problem that the traditional method cannot distinguish between "inherent pollution" and "interaction pollution". The pollution resonance effect is often caused by the synergistic interaction of multiple metals, for example, the release of chromium and nickel by the previous equipment, which interacts with the easily adsorbed lead of the subsequent equipment, and the sum Covering all pairs of metal elements involved in the interaction, reflecting the integrity of "multi-element synergistic amplification", avoiding the omission caused by single-element pair analysis. The metal ions released by the previous equipment will be lost due to volatilization, sedimentation, adsorption on the container wall, etc. during the transfer process, and their concentration decays exponentially with time, the exponential term in the formula Precise quantification of the weakening effect of time on the interaction effect, solving the defect of ignoring the "time dimension influence" in the traditional method. The final manifestation of the pollution resonance effect is the amplification of the concentration of the affected metal, and the order sensitivity index of the affected metal Directly reflects its response strength to the equipment order, The larger the concentration fluctuates with the equipment order, the more significant the interaction effect can be amplified when the affected party is sensitive to the order. The As a weight term, it accurately captures the core law that "order sensitivity is the key carrier of resonance amplification", avoiding logical misplacement caused by misusing the sensitivity index of the affected metal. The interaction enhancement factor calculation formula restores the formation path of the pollution resonance effect through "baseline correction-increment extraction-multi-element summation-time decay-sensitivity weighting" hierarchical design. This formula breaks through the limitations of traditional "linear superposition", coupling "metal affinity, time, and order" three independent influencing factors into a unified index, making the deviation caused by resonance effect in low-concentration samples can be accurately quantified, solving the core calibration problem of "micro-cumulative amplification into significant deviation". In the formula Binds with the specific equipment transfer time, Binds with the equipment release / adsorption spectrum, Binds with the element sensitivity characteristics, so that the interaction enhancement factor can adapt to personalized scenarios of different equipment combinations and different processing rhythms, avoiding distortion caused by "one-size-fits-all" calibration.

[0141] Step S33, based on the interaction enhancement factor of the continuous equipment pair, a three-level pollution amplification model is constructed to calculate the cumulative interaction coefficient;

[0142] Step S33 aims to construct a three-level pollution amplification model based on the topological structure of the equipment transfer chain, and calculate the cumulative interaction coefficient representing the degree of pollution amplification in the whole process by integrating the direct and indirect interaction effects between equipment.Figure 4 As shown, the three-stage pollution amplification model is a hierarchical analysis framework for analyzing the transmission path of pollution effect. The first stage is the internal pollution accumulation of a single device, i.e., the element-specific total pollution contribution calculated in step S10, which refers to the cumulative amount of heavy metal pollution generated by a single device due to wear, adsorption, etc. in the sample contact path, and is the basis carrier for subsequent interactive amplification effect. The second stage is the direct interaction between adjacent devices, i.e., the interactive enhancement factor calculated in step S32, which refers to the direct interactive amplification effect caused by the adsorption characteristics of the subsequent device on the released metal of the previous device in the continuous device pair. The third stage is the indirect interaction transmission across devices, which refers to the superimposed influence of the interactive effect of the previous device pair on the subsequent device pair through the pollution transmission of the intermediate device, and the essence is the cascade conduction and superimposed amplification of the interactive effect. The cumulative interaction coefficient is a parameter specific to each heavy metal element, and its core function is to represent the total pollution amplification degree under the combined action of “direct interaction + indirect transmission” in the whole process.

[0143] The specific implementation process is as follows: First, build a three-stage pollution amplification model based on the topological structure of the device transfer chain. The topological structure needs to completely restore the connection order and transmission relationship of the sample processing devices, for example, when the device transfer chain is “ball mill → rotary evaporator → ultrasonic cleaner”, the topological structure presents a linear connection of “ball mill (node 1) — rotary evaporator (node 2) — ultrasonic cleaner (node 3)”, and the continuous device pairs are “node 1→ node 2” and “node 2→ node 3”. It is clear that the internal pollution accumulation of each node (device) in the first stage has been quantified in step S10, the second stage focuses on the direct interaction between two continuous device pairs in the device transfer chain, i.e., the interactive enhancement factor of the continuous device pair “node 1→ node 2” and the interactive enhancement factor of the continuous device pair “node 2→ node 3”, and the third stage focuses on the interactive effect represented by the interactive enhancement factor of the continuous device pair “node 1→ node 2”, which is transmitted to the continuous device pair “node 2→ node 3” through the rotary evaporator as an intermediate node, and produces superimposed amplification with the interactive effect represented by the interactive enhancement factor of the device pair.

[0144] Substitute the interactive enhancement factors of all continuous device pairs into the cumulative interaction coefficient calculation formula:

[0145]

[0146] Get the cumulative interaction coefficient, where:

[0147] : cumulative interaction coefficient, representing the total pollution amplification degree of “adjacent device direct interaction + cross-device indirect transmission” in the whole process.

[0148] : multiplication operator, where m is the total number of continuous device pairs in the device transfer chain, Index of the continuous device pair, =1 represents the first continuous device pair, and the superscript m represents the mth continuous device pair, covering all continuous device pairs in the device transfer chain.

[0149] : the interaction enhancement factor of the mth continuous device pair, representing the direct interaction amplification effect of the front device releasing metal and the rear device adsorbing metal in the device pair.

[0150] : chain length amplification coefficient, quantifying the strengthening effect of the increase in the number of devices on the indirect interaction transmission between devices. The more the number of devices, the more complex the intermediate transmission link, and the more significant the superposition effect of indirect interaction. The determination method is as follows: select device transfer chains with different node numbers (such as 3, 4, 5, and 6), respectively detect the heavy metal concentrations of the same standard sample after treatment, calculate the actual pollution amplification multiples under different chain lengths, establish the correlation curve of “amplification multiple-chain length”, and determine the value range of by fitting, with an exemplary value of 0.1-0.2.

[0151] : the number of nodes in the device transfer chain, i.e. the total number of devices involved in the sample pretreatment process.

[0152] : the maximum number of pretreatment devices allowed by the system, which is determined according to the range of the number of devices commonly used in the actual pretreatment process. By investigating the typical pretreatment process of soil heavy metal detection in the industry, the upper limit of the number of devices covering more than 95% of the scenarios is selected as N max , with an exemplary value of 6.

[0153] In the formula, the product of the interaction enhancement factors of all continuous device pairs in the device transfer chain is calculated, which physically represents the cumulative effect of the direct interaction of adjacent device pairs, and essentially quantifies the third-level “indirect interaction transmission between devices”. The interaction amplification effect of the previous device pair will become an “additional pollution source” for the subsequent interaction through the pollution residues and transmission of the intermediate devices, resulting in the amplification effect of the subsequent γ value being built on the basis of “basic pollution + previous interaction amplification pollution”. The multiplication operation accurately restores the cascading transmission rule of “previous amplification results as the basis for subsequent amplification”. The design logic of is as follows: when N=N max , the amplification effect of chain length on indirect transmission reaches a peak; when N is small, the amplification effect is weak, which conforms to the objective law that the more the number of devices, the more significant the superposition of indirect interaction.

[0154] ​​​The prior art calculates the pollution contribution by linear superposition, which neither considers the direct interaction of adjacent equipment nor the indirect transmission and chain length effect across equipment, resulting in the pollution amplification effect of long process samples being seriously underestimated. Step S33 builds a three-level model, clarifies the complete pollution path of "basic pollution-direct interaction-indirect transmission", and solves the problem of incomplete analysis of the transmission path of pollution effect in traditional models; through the combination of product and chain length correction in the formula, the cascade amplification and chain length reinforcement of indirect interaction are accurately quantified, and the technical problem of "indirect interaction and chain length effect cannot be quantified" is solved.

[0155] In step S34, based on the cumulative interaction coefficient, the dynamic interference quantity of each heavy metal element is calculated respectively, and the final calibration value of each heavy metal element in the soil sample to be calibrated is obtained by deducting the dynamic interference quantity from the preliminary calibration value.

[0156] The dynamic interference quantity refers to the additional pollution contribution caused by the interaction between metals and the sequence of equipment. For each heavy metal element q, the dynamic interference quantity is calculated as follows: ΔC q,dynamic =C q,preliminary ×(Γ q -1)×f q , wherein C q,preliminary is the preliminary calibration value of element q, Γ q is the cumulative interaction coefficient of element q, and f q is the element-specific calibration factor.

[0157] The determination method of the element-specific calibration factor f q is as follows: if element q is a sequence-sensitive element, f q =1.0, because such elements respond strongly to interaction effects; if element q is a non-sequence-sensitive element, f q =0.5, because such elements respond weakly to interaction effects; and if element q does not participate in any interaction, f q =0, because such elements have no additional pollution caused by interaction. The element not participating in any interaction refers to an element that has neither a non-zero interaction influence strength value as an influencing party nor a non-zero interaction influence strength value as an influenced party in the metal ion affinity matrix of the pollution interaction characteristic library. Finally, the final calibration value is calculated by the formula C q,secondary =C q,preliminary -ΔC q,dynamic , and the calibration is completed, wherein C q,secondary is the final calibration value of element q.

[0158] The prior art uses a unified calibration standard for all elements without considering the response difference of elements, resulting in large calibration deviation of sensitive elements. Step S34 solves this problem by differentiating f q , so that the dynamic interference quantity calculation adapts to the inherent characteristics of the elements. In the formula, (Γq -1) stripping the baseline of non-interaction, leaving only the interaction-induced amplification part, with f q Combined with the implementation of element differentiation correction, it logically meets the actual needs of "sensitive element correction and non-sensitive element correction". This step cooperates with steps S13 and S23: the preliminary calibration value of step S13 provides a basic value for calculation, and the sequential sensitive element list of step S23 provides f q The setting provides the basis, and the combination of the three realizes the closed loop of "spatial and temporal non-uniformity calibration" and "non-linear interaction calibration". If this step is missing, the cumulative interaction coefficient cannot be converted into a subtractable interference quantity, all previous analyses about interaction effects cannot be implemented, and the systematic deviation caused by "pollution resonance effect" cannot be eliminated.

[0159] Step S30 solves the problem of "pollution resonance effect" caused by the "micro cumulative effect" of the pretreatment device through the technical path of "chain construction-factor calculation-model amplification-differentiated subtraction", that is, the deviation amplification problem caused by the superposition of intermetallic non-linear interaction and sequential sensitivity of the device. By clearly defining the interaction scene through the device transfer chain, combining the metal ion affinity matrix and the sequential sensitivity index, the abstract "pollution resonance effect" is converted into quantifiable γ and Γ parameters, solving the defect that traditional methods cannot identify non-linear amplification. From direct interaction between single devices to cascading transmission among multiple devices, from element response difference to time decay effect, it covers the interference factors of the entire pretreatment process, solves the problem that traditional static mean subtraction method cannot adapt to dynamic processing scenarios, and significantly improves the calibration accuracy. Step S30 cooperates with the preliminary calibration of step S10 and the construction of the interaction feature library of step S20. Step S10 removes the spatiotemporal non-uniformity interference, step S20 provides interaction rule support, and step S30 completes the subtraction of non-linear interaction interference. The combination of the three upgrades the entire calibration scheme from "single factor correction" to "multi-dimensional system calibration", completely solving the core problem of misjudgment of clean soil. If step S30 is missing, the preliminary calibration of step S10 can only eliminate the linear cumulative pollution of a single device, and cannot handle the non-linear interaction amplification between devices. The pollution interaction feature library of step S20 cannot be converted into an actual calibration tool, and the "systematic deviation amplification" problem of the entire scheme is still not solved, and the risk of misjudging clean soil as slightly contaminated continues to exist.

[0160] Example 2:

[0161] This embodiment is based on example 1 and provides an adaptive calibration system for soil heavy metal detection data, as shown in Figure 5 , which includes

[0162] The pollution contribution calculation module is configured to construct a pollution fingerprint database of the pretreatment equipment, establish a sample-equipment contact matrix, obtain original detection values of each heavy metal element in the to-be-calibrated soil sample, and calculate element-specific total pollution contribution based on the pollution fingerprint database and the sample-equipment contact matrix.

[0163] The preliminary calibration module is configured to deduct the element-specific total pollution contribution from the original detection values of each heavy metal element in the to-be-calibrated soil sample to obtain preliminary calibration values of each heavy metal element in the to-be-calibrated soil sample.

[0164] The pollution interaction quantification module is configured to draw a pollution interaction curve through a gradient concentration experiment, design an orthogonal experiment scheme, and calculate a sequence sensitivity index; construct a metal ion affinity matrix based on the pollution interaction curve and the sequence sensitivity index; and integrate the metal ion affinity matrix, the sequence sensitivity index, and the pollution interaction curve to establish a pollution interaction feature library.

[0165] The final calibration module is configured to extract an equipment contact sequence of the to-be-calibrated soil sample from the sample-equipment contact matrix, calculate a dynamic interference quantity of each heavy metal element based on the pollution interaction feature library and the equipment contact sequence of the to-be-calibrated soil sample, and deduct the dynamic interference quantity from the preliminary calibration values to obtain final calibration values of each heavy metal element in the to-be-calibrated soil sample.

[0166] In the pollution interaction quantification module, the method for calculating the sequence sensitivity index comprises:

[0167] designing an orthogonal experiment scheme, changing the use sequence of the pretreatment equipment, generating a plurality of equipment combination schemes, processing the same standard soil sample by using each equipment combination scheme, detecting the standard soil samples processed by different equipment combination schemes, recording the detection concentrations of each heavy metal, establishing an element-scheme-concentration three-dimensional data table, extracting the concentration set of each heavy metal element in all equipment combination schemes from the element-scheme-concentration three-dimensional data table, calculating the highest concentration, the lowest concentration, and the average concentration of the current heavy metal element, and calculating the sequence sensitivity index of the current heavy metal element according to the highest concentration, the lowest concentration, and the average concentration of the current heavy metal element.

[0168] In the final calibration module, the method for calculating the dynamic interference quantity of each heavy metal element comprises:

[0169] based on the equipment contact sequence, identifying a continuous equipment pair, and constructing an equipment transfer chain;

[0170] based on the equipment transfer chain and the pollution interaction feature library, calculating an interaction enhancement factor of each continuous equipment pair;

[0171] based on the interaction enhancement factor of the continuous equipment pair, constructing a three-level pollution amplification model, and calculating a cumulative interaction coefficient.

[0172] Based on the cumulative interaction coefficient, the dynamic interference quantity is calculated for each heavy metal element.

[0173] The methods and systems of this application can be implemented in a number of ways. For example, the methods and systems of this application can be implemented via software, hardware, firmware, or any combination of software, hardware, and / or firmware. The above-described order of steps for the methods is merely illustrative, and the steps of the methods of this application are not limited to the specific order described above unless otherwise specifically stated.

[0174] In addition, parts of the above technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of corresponding technical solutions in the prior art are not described in detail to avoid excessive repetition.

[0175] The specific embodiments described above are further explained with reference to the accompanying drawings. The purpose of this description is to illustrate the inventive aspects of this application and not to limit the scope of the application. Any modifications made to the above-described embodiments should be considered within the scope of the present application.

Claims

1. An adaptive calibration method for soil heavy metal detection data, characterized in that, The method includes: Pollutant samples were collected from different spatial locations of multiple pretreatment devices, and the pollutant samples were tested to obtain pollutant concentration data sets from different pretreatment devices. The sampling time was recorded, and the pollutant concentration data sets from different pretreatment devices were integrated to construct a pollution fingerprint database of the pretreatment devices. A unique identification code is assigned to each soil sample. A position sensor is installed in the pretreatment equipment to establish an equipment position sensing network. When a soil sample with a unique identification code enters the pretreatment equipment, the equipment position sensing network records the pretreatment equipment number, the soil sample's position coordinates in the pretreatment equipment, and the dwell time at each position coordinate. Based on the dwell time of the soil sample at each position coordinate in the pretreatment equipment, the contact intensity coefficient at each position coordinate is calculated. A sample-equipment contact matrix is ​​constructed with the soil sample's unique identification code as the row and the position coordinates in the pretreatment equipment as the column. The contact intensity coefficient, dwell time, and equipment number of each soil sample at different position coordinates in the pretreatment equipment are filled into the corresponding positions in the sample-equipment contact matrix. The original detection values ​​of each heavy metal element in the soil sample to be calibrated are obtained. Based on the pollution fingerprint database and the sample-equipment contact matrix, the element-specific total pollution contribution is calculated. The element-specific total pollution contribution is subtracted from the original detection values ​​of each heavy metal element in the soil sample to be calibrated to obtain the preliminary calibration value of each heavy metal element in the soil sample to be calibrated. Pollution interaction curves were plotted through gradient concentration experiments, orthogonal experimental schemes were designed, and order sensitivity indices were calculated. Based on the pollution interaction curves and order sensitivity indices, metal ion affinity matrices were constructed. The metal ion affinity matrix, order sensitivity indices, and pollution interaction curves were integrated to establish a pollution interaction feature library. The equipment contact sequence of the soil sample to be calibrated is extracted from the sample-equipment contact matrix. Based on the pollution interaction feature library and the equipment contact sequence of the soil sample to be calibrated, the dynamic interference amount of each heavy metal element is calculated. The dynamic interference amount is subtracted from the preliminary calibration value to obtain the final calibration value of each heavy metal element in the soil sample to be calibrated.

2. The adaptive calibration method for soil heavy metal detection data according to claim 1, characterized in that, The method for calculating the contact strength coefficient at each location coordinate includes: If the residence time t of the soil sample at any coordinate in the pretreatment equipment is... stay The contamination migration threshold t of the corresponding pretreatment equipment is greater than thr Then, the contact intensity coefficient at the corresponding location coordinates is calculated based on the residence time and pollution migration threshold; if t stay Less than or equal to t thr , then I=0.

3. The adaptive calibration method for soil heavy metal detection data according to claim 2, characterized in that, The calculation method for the element-specific total pollution contribution includes: The contamination migration path of the soil sample to be calibrated can be obtained by querying the sample-equipment contact matrix based on the unique identification code of the soil sample to be calibrated. For each contact point in the contamination migration path of the soil sample to be calibrated, the pollution contribution baseline value at the corresponding location is extracted from the contamination fingerprint database, and the pollution contribution baseline value is corrected for time to obtain the corrected pollution contribution value. The total pollution contribution is obtained by traversing all contact points in the pollution migration path of the soil sample to be calibrated and summing the corrected pollution contribution values ​​of each contact point. If the total pollution contribution is greater than 5% of the original detection value, then enter the deep calibration mode to obtain the element-specific total pollution contribution.

4. The adaptive calibration method for soil heavy metal detection data according to claim 3, characterized in that, The method for time-correcting the pollution contribution baseline value includes: Determine the interval Δt between the contact time and the sampling time at any contact point in the contamination migration path of the soil sample to be calibrated. When the interval Δt is greater than the effective period t... valid At that time, the pollution contribution baseline value of the corresponding contact point is corrected for time.

5. The adaptive calibration method for soil heavy metal detection data according to claim 4, characterized in that, The method for plotting the pollution interaction curve includes: A series of interactive source metal standard soils with gradient concentrations were prepared, and each concentration of interactive source metal standard soil was divided into two groups and treated with different equipment combinations. Heavy metal full-spectrum analysis was performed on the standard soil samples treated with two different equipment combinations to calculate the additional release of the interacting metals. The pollution interaction curve was plotted with the concentration of the interacting source metals on the horizontal axis and the additional release of the interacting metals on the vertical axis.

6. The adaptive calibration method for soil heavy metal detection data according to claim 5, characterized in that, The method for calculating the order sensitivity index includes: In the orthogonal experimental scheme, the order of use of the pretreatment equipment is changed to generate multiple equipment combination schemes. The same standard soil sample is treated with each equipment combination scheme. The standard soil samples treated with different equipment combination schemes are tested, the detection concentration of each heavy metal is recorded, and a three-dimensional data table of element-scheme-concentration is established. Extract the concentration set of each heavy metal element in all equipment combination schemes from the element-scheme-concentration three-dimensional data table, calculate the highest, lowest and average concentrations of the current heavy metal element, and calculate the order sensitivity index of the current heavy metal element based on the highest, lowest and average concentrations.

7. The adaptive calibration method for soil heavy metal detection data according to claim 6, characterized in that, The method for constructing the metal ion affinity matrix includes: Based on the pollution interaction curve, calculate the baseline value of the interaction effect intensity; Determine whether there is intermetallic interaction between the source metal and the interacted metal based on the pollution interaction curve; The metal that interacts with another metal is defined as the influencing metal, and the metal that is interacted with is defined as the affected metal. It is determined whether the influencing and affected metals are order-sensitive elements. If both the influencing and affected metals are order-sensitive elements, the basic value of the interaction influence intensity is corrected according to the order sensitivity index of the affected metal to obtain the final interaction influence intensity value; otherwise, the basic value of the interaction influence intensity is directly used as the final interaction influence intensity value. Based on the final interaction strength values, a metal ion affinity matrix is ​​constructed.

8. An adaptive calibration system for soil heavy metal detection data, used to implement the adaptive calibration method for soil heavy metal detection data according to any one of claims 1-7, characterized in that, The system includes: Pollution contribution calculation module: used to construct a pollution fingerprint spectrum database of the pretreatment equipment and establish a sample-equipment contact matrix, obtain the original detection value of each heavy metal element in the soil sample to be calibrated, and calculate the element-specific total pollution contribution based on the pollution fingerprint spectrum database and the sample-equipment contact matrix. Preliminary calibration module: used to subtract the element-specific total pollution contribution from the original detection value of each heavy metal element in the soil sample to be calibrated to obtain the preliminary calibration value of each heavy metal element in the soil sample to be calibrated; Pollution Interaction Quantification Module: Used to plot pollution interaction curves through gradient concentration experiments, design orthogonal experimental schemes, and calculate the order sensitivity index; construct a metal ion affinity matrix based on the pollution interaction curve and the order sensitivity index; and integrate the metal ion affinity matrix, the order sensitivity index, and the pollution interaction curve to establish a pollution interaction feature library. Final calibration module: used to extract the device contact sequence of the soil sample to be calibrated from the sample-device contact matrix, calculate the dynamic interference amount of each heavy metal element based on the pollution interaction feature library and the device contact sequence of the soil sample to be calibrated, and subtract the dynamic interference amount from the preliminary calibration value to obtain the final calibration value of each heavy metal element in the soil sample to be calibrated.

Images