Self-adaptive sampling method and system for multi-parameter sensing of river sediment

Abstract

The invention discloses a river sediment multi-parameter sensing self-adaptive sampling method and system, and relates to the technical field of environment monitoring, and the method comprises the following steps: S1, synchronously collecting the pH value, electrochemical potential, heavy metal binding state proportion, water body flow velocity and hydrodynamic parameters of sediment through a sensing device, and establishing a multi-parameter response baseline; and S2, based on the multi-parameter response baseline, collecting monitoring data in real time and executing differential analysis, and identifying a high-risk time period when the pH value change rate exceeds a threshold value and the heavy metal binding state proportion fluctuates. By constructing a multi-parameter response baseline and a linkage recognition mechanism, early sensing and accurate sampling of heavy metal release are realized, a closed-loop response system is constructed by combining chemical form analysis, self-correction and data feedback optimization recognition model, and the accuracy and intelligent level of pollution monitoring are improved.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, specifically to an adaptive sampling method and system for multi-parameter sensing of river sediments. Background Technology

[0002] Adaptive sampling with multi-parameter sensing of river sediments refers to the real-time sensing and analysis of multi-dimensional characteristic parameters of sediments (such as particle size distribution, organic matter content, heavy metal concentration, water content, temperature, and conductivity) during river sediment monitoring. This is achieved by deploying multiple types of sensing units and data acquisition devices, and by combining this data with external hydrological conditions such as water flow velocity, direction, and water level changes to dynamically assess the changing trends of the current sediment state. Based on the sensed multi-parameter data, the system adjusts the sampling time interval, spatial location, sampling depth, and sampling frequency in real time. This enables focused monitoring of key periods and abnormal areas, avoiding data redundancy and information loss caused by blind or fixed-pattern sampling. Thus, while ensuring data accuracy and representativeness, it improves sampling efficiency and resource utilization, providing precise and dynamic monitoring support for river pollution source tracing, ecological restoration, and water environment management.

[0003] Existing technologies have the following shortcomings: During river sediment monitoring, some heavy metals in sediments typically exist in stable forms bound to organic matter, sulfides, or oxides, making them difficult to directly enter the water. However, when the pH of river water changes drastically in a short period, such as due to acidic rainwater runoff, sudden discharge of upstream industrial wastewater, or mine tailings leakage, the chemical balance within the sediments is disrupted. This causes the previously stable heavy metal binding states to dissociate instantaneously and transform into soluble states. These soluble heavy metals rapidly diffuse and migrate under the influence of water flow, resulting in a peak increase in pollutant concentration in the water within a very short time. Due to the rapid rate and short duration of this process, existing adaptive sampling systems often struggle to promptly sense and adjust sampling strategies in the early stages of sudden pH parameter changes. This leads to delays or even complete omissions in the detection of such instantaneous high-concentration heavy metal events, affecting the accurate capture of pollution peaks and the accurate assessment of pollution risks.

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

[0005] The purpose of this invention is to provide an adaptive sampling method and system for multi-parameter sensing of river sediments to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an adaptive sampling method for multi-parameter sensing of river sediments, comprising the following steps:

[0007] S1, by synchronously collecting the pH, electrochemical potential, heavy metal binding state ratio of sediments, as well as water flow velocity and hydrodynamic parameters through sensing devices, to establish a multi-parameter response baseline;

[0008] S2, based on a multi-parameter response baseline, collects monitoring data in real time and performs differential analysis to identify high-risk periods when the rate of change in pH exceeds the threshold and the proportion of heavy metal bound states fluctuates.

[0009] S3, after identifying high-risk periods, triggers a priority sampling strategy, adjusts the spatial position and depth of the sampling device, and collects samples from layers with active heavy metal release and high hydrodynamic areas.

[0010] S4 performs rapid chemical speciation analysis on the collected samples to obtain the proportion of heavy metal bound state and the concentration of soluble state. Based on the comparison results, the risk judgment threshold is corrected to achieve system self-calibration.

[0011] S5, based on threshold correction, redistributes sampling frequencies according to the hydrodynamic and chemical characteristics of each monitoring area, prioritizing areas with similar disturbance characteristics to achieve pollution migration path coverage;

[0012] S6 inputs the full-process monitoring data and response records into the adaptive optimization module, updates the multi-parameter response baseline and differential identification model, and constructs a closed-loop response mechanism with dynamic control capabilities.

[0013] Preferably, step S1 includes:

[0014] Multiple monitoring points were set up at representative locations in the target monitoring section of the river channel. Each monitoring point was equipped with multiple cross-sectional and vertical sampling points, and pH sensors, electrochemical potential electrodes, conductivity sensors, flow velocity sensors and water depth sensors were installed.

[0015] The proportions of acid-soluble, reducible, oxidizable, and residual heavy metals in sediments were obtained by using a water sample extraction device combined with sequential extraction and quantitative analysis.

[0016] The data collection frequency was set to once every 15 minutes, with a continuous collection period of no less than 14 days, and the collected data covered typical hydrological processes.

[0017] Parametric analysis is performed on structured sample data to extract the mean, standard deviation and variation under natural steady-state conditions, establish a response baseline with clear numerical boundaries, and revise and update it monthly.

[0018] Preferably, step S2 includes:

[0019] Monitoring data including pH, electrochemical potential, mass concentration ratio of various heavy metal forms, water flow velocity, water depth, water temperature, conductivity, dissolved oxygen concentration and turbidity are acquired at a time frequency of 15 minutes.

[0020] The collected data is preprocessed and compared item by item with the multi-parameter response baseline;

[0021] When the rate of change in pH exceeds 0.5 units, the electrochemical potential decreases by more than 30 millivolts, the proportion of heavy metals in reducible or acid-soluble form increases by more than 10%, and the water flow velocity increases significantly, it is identified as a potentially high-risk period.

[0022] Sliding window trend analysis is performed on high-risk states over a continuous period of time to identify high-risk phases and record the basis for the judgment.

[0023] Preferably, the specific steps for performing sliding window trend analysis on high-risk states over a continuous time period to identify high-risk phases and record the criteria for judgment are as follows:

[0024] Set a sliding analysis window with a time length of 30 minutes to perform a coverage scan of continuously sampled data;

[0025] Within each sliding window, record at least two instances of the combined criteria of pH change rate, electrochemical potential decrease, heavy metal occurrence speciation change, and water flow enhancement are collected.

[0026] When the judgment parameters show a continuous increasing trend within the window time, it is confirmed as a high-risk stage of real heavy metal release, and the time interval of risk judgment, sampling profile number, key parameter change value and recommended response measures are recorded.

[0027] Preferably, step S3 includes:

[0028] Based on the high-risk assessment results, extract the trigger parameter type, sampling profile number, identification time point, and corresponding sampling depth;

[0029] High-risk profiles were sorted according to risk level, and multiple sampling points were set up laterally along the river channel with a fixed vertical sampling depth.

[0030] Sediment samples were extracted using a negative pressure sampling chamber and labeled. The samples were transported in a 4°C refrigerated environment.

[0031] When the flow rate and particulate matter concentration reach the disturbance conditions, the sampling profile is expanded along the downstream main axis and the sampling operation is performed simultaneously.

[0032] After sampling is completed, a sampling execution log containing time, coordinates, depth, and triggering basis is created and archived.

[0033] Preferably, step S4 includes:

[0034] Sediment samples collected from high-risk profiles were numbered, checked, and pretreated. Acid-soluble, reducible, oxidizable, and residual heavy metals were extracted using a sequential extraction method, and their concentrations were determined using inductively coupled plasma mass spectrometry.

[0035] Extract the proportion of acid-soluble and reducible metal concentrations and compare it with the risk identification and judgment results;

[0036] Based on the comparison deviation, the risk assessment threshold is adjusted by a fixed percentage to form a risk threshold correction file;

[0037] The revised parameters are applied to the next round of real-time monitoring to determine high-risk states, forming a closed-loop feedback process.

[0038] Preferably, step S5 includes:

[0039] Collect hydrodynamic and chemical response parameters from all monitoring profile points and calculate the pollution sensitivity index (CSI) value.

[0040] Based on the CSI value, the monitoring area is divided into high-sensitivity zone, medium-sensitivity zone and low-sensitivity zone, and sampling cycles of 15 minutes, 30 minutes and 60 minutes are set respectively.

[0041] The actual migration distance of the pollution downstream is calculated based on the current flow rate and sampling interval. High-frequency sampling points are then deployed along the main flow direction within this distance range to ensure full coverage and continuous tracking of the pollution migration path.

[0042] Preferably, step S6 includes:

[0043] The data on pH, electrochemical potential, flow rate, water level, temperature, dissolved oxygen, conductivity, and concentration of various forms of heavy metals during the current monitoring period, along with the sampling and adjustment records, will be archived together.

[0044] The maximum, minimum, average, and standard deviation of non-high-risk status data within 48 hours are calculated to form a new multi-parameter response baseline.

[0045] Based on the new baseline adjustment difference identification criteria, the trigger thresholds of each judgment parameter are numerically corrected and an identification criterion correction record table is formed;

[0046] The updated multi-parameter response baseline and identification criteria are used as the basis for judgment and input into the next round of monitoring process to build a dynamically updated closed-loop response mechanism.

[0047] Preferably, the specific steps for numerically correcting the trigger thresholds of each judgment parameter and forming an identification criterion correction record table based on the new baseline adjustment difference identification criterion are as follows:

[0048] Statistical analysis was performed on the variation range of each monitoring parameter within the normal fluctuation range. The trigger thresholds for pH value, electrochemical potential, and reducible heavy metal concentration were reset. The parameter names before and after the correction, the original threshold, the corrected threshold, the correction time, and the data on which they were based were recorded in the identification criterion correction record table. This record was used to replace the original judgment criteria and guide the next round of risk identification and judgment process.

[0049] An adaptive sampling system for multi-parameter sensing of channel sediments includes a multi-parameter baseline construction module, a risk identification and differential judgment module, a priority sampling control module, a rapid morphological analysis and threshold correction module, a regional frequency scheduling module, and a closed-loop optimization and model update module.

[0050] The multi-parameter baseline construction module synchronously collects sediment pH, electrochemical potential, heavy metal bound state ratio, water flow velocity and hydrodynamic parameters through sensing devices to establish a multi-parameter response baseline.

[0051] The risk identification and differential judgment module, based on a multi-parameter response baseline, collects monitoring data in real time and performs differential analysis to identify high-risk periods where the rate of change in pH exceeds the threshold and the proportion of heavy metal bound states fluctuates.

[0052] The priority sampling control module identifies high-risk periods and triggers a priority sampling strategy to adjust the spatial position and depth of the sampling device to collect samples from layers with active heavy metal release and high hydrodynamic regions.

[0053] The rapid morphology analysis and threshold correction module performs rapid chemical morphology analysis on the collected samples to obtain the proportion of heavy metal bound states and the concentration of soluble states. Based on the comparison results, it corrects the risk judgment threshold to achieve system self-calibration.

[0054] The regional frequency scheduling module, based on threshold correction, reallocates sampling frequencies according to the hydrodynamic and chemical characteristics of each monitoring area, prioritizing scheduling to areas with similar disturbance characteristics to achieve pollution migration path coverage;

[0055] The closed-loop optimization and model update module inputs the monitoring data and response records of the entire process into the adaptive optimization module to update the multi-parameter response baseline and differential identification model, and build a closed-loop response mechanism with dynamic control capabilities.

[0056] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0057] This invention significantly enhances the early detection capability of abrupt pH changes by establishing a multi-parameter response baseline in the initial monitoring phase and introducing a linkage identification mechanism involving pH, electrochemical potential, heavy metal binding state ratio, and hydrodynamic characteristics during sampling. Building upon this, the system dynamically adjusts sampling location and frequency, precisely identifies high-risk layers, rapidly conducts chemical speciation analysis, and incorporates self-calibration and threshold update mechanisms to ensure rapid response and real-time correction of judgment criteria during critical periods of heavy metal release, effectively avoiding data omissions and misjudgments. Finally, through full-process data feedback and dynamic optimization of the differential identification model, a closed-loop response mechanism with self-learning and adaptive capabilities is constructed, comprehensively improving the accuracy, reliability, and intelligence of heavy metal pollution monitoring in river sediments. This provides solid data support and decision-making basis for pollution early warning, source tracing, and ecological restoration. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0059] Figure 1 This is a flowchart of the adaptive sampling method for multi-parameter sensing of river sediments according to the present invention.

[0060] Figure 2 This is a schematic diagram of the module of the adaptive sampling system for multi-parameter sensing of river sediments according to the present invention. Detailed Implementation

[0061] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0062] This invention provides, for example Figure 1 The adaptive sampling method for multi-parameter sensing of channel sediments, as shown, includes the following steps:

[0063] S1. In the early stage of river monitoring, the pH value, electrochemical potential, heavy metal binding state ratio of sediments, as well as the flow velocity and hydrodynamic parameters of river water are collected simultaneously through sensing devices to construct a multi-parameter response baseline, which is used to characterize the coupling relationship between sediment stability characteristics and hydrodynamic characteristics under normal conditions.

[0064] The adaptive sampling method for multi-parameter sensing of river sediments constructs a multi-parameter response baseline in the early stages of river monitoring through the following steps:

[0065] Multiple monitoring points were deployed in the selected target monitoring sections of the river channel. Specifically, representative straight sections, confluences of slow and fast currents, tributary inlets, or areas with frequent human activity were selected as monitoring points. At each monitoring point, 3 to 5 sampling profiles were established along the river cross-section, with at least 3 vertical sampling points along the water depth direction in each profile, extending 20 to 50 cm below the water-sediment interface. Integrated sensing equipment was installed at each sampling point, including a miniature pH electrode for measuring pH, a silver / silver chloride reference electrode for measuring redox potential, a conductivity sensor, a flow velocity sensor, and a water depth sensor. The device for detecting the proportion of bound heavy metals consisted of an online water sample extraction unit and a stepwise extraction and analysis unit. Sequential extraction was used to separate acid-soluble, reducible, oxidizable, and residual heavy metal components, followed by quantitative analysis using colorimetry or inductively coupled plasma mass spectrometry. All sensors were connected to the shore-based data acquisition platform via fiber optic communication or a low-power wireless network for synchronous data acquisition and uploading. The data collection frequency is set to once every 15 minutes, and the continuous operation is no less than 14 days. The sampling period covers typical hydrological events such as drought days, rainy days, and high and low tides to ensure that the collected data is representative and has periodic characteristics.

[0066] The collected raw data underwent structured processing, extracting pH, potential, water flow velocity, water level change, concentration ratios of four heavy metal forms in sediments, and auxiliary parameters such as water temperature and conductivity. During processing, outliers were eliminated by setting reasonable quality control standards, such as removing isolated readings with pH fluctuations exceeding one unit without accompanying water level changes, or eliminating instantaneous potential jumps caused by sensor malfunctions. Subsequently, all data were paired grouped according to sampling time and spatial location to form a multi-parameter monitoring sample set. For each sample set, the correlation between different parameters was analyzed, such as whether the redox potential of sediments decreased synchronously with increasing water flow velocity, or whether there were significant changes in pH and the proportion of reducible heavy metals when the water level decreased. By calculating the linear regression coefficients, trend directions, and response time lags between different parameters, the stable response relationships of each chemical parameter under natural hydrodynamic driving were clarified.

[0067] Based on the aforementioned sample set, characteristic response patterns representing sediments in an undisturbed state are extracted. Specifically, samples from periods of stable water flow, minimal wind speed fluctuations, no human intervention, and no pollution inflow are selected. The average value, standard deviation, and relative variation of each parameter in these samples are extracted to form a set of benchmark parameters considered to possess "natural steady-state characteristics." For example, samples with pH values ​​between 6.8 and 7.4, electrochemical potentials between +150 and +250 mV, the proportion of reducible heavy metals maintained between 30% and 40% of the total, and water flow velocities less than 0.2 m / s can be selected as a steady-state reference set. Based on this set, a set of response baselines with clearly defined numerical boundaries and relative relationships is established. For example, the response baseline can specify that when the flow velocity increases by 0.1 m / s, the pH decrease should not exceed 0.2 units, the electrochemical potential decrease should not exceed 30 mV, and the change in the proportion of bound heavy metals should not exceed 5%. In this way, complex chemical behaviors are transformed into detectable, calculable, and comparable benchmark linear indicators.

[0068] The constructed multi-parameter response baseline serves as a reference for the monitoring system to determine whether the current state deviates from the natural steady state during the identification of sudden disturbances. In subsequent monitoring, if the data collected within a certain period simultaneously meets the following characteristics: pH change rate exceeding 0.5 units / hour, electrochemical potential decrease exceeding 50 mV, and the proportion of reducible heavy metals increasing by more than 10%, accompanied by a water flow velocity increase exceeding 0.15 m / s, then that period is determined to be the initial abnormal state, and the subsequent high-risk identification and sampling response process is initiated. To adapt to seasonal changes and long-term trend variations in the river channel, the response baseline is updated every month. The update process includes recalculating stable period characteristic samples, correcting parameter ranges, and adjusting comparison thresholds to ensure that the baseline always closely reflects actual environmental conditions and does not become invalid due to background drift.

[0069] This step serves to provide a stable and reliable reference for subsequent identification of potential heavy metal release risks in river channels, forming the foundation and prerequisite for the entire adaptive sampling method. In river environments, heavy metals in sediments are typically in a relatively stable bound state, and their occurrence is closely related to the surrounding chemical environment (acidity / alkalinity, electrochemical potential) and hydrodynamic conditions (flow velocity, water depth, intensity of hydrodynamic disturbance). Under normal conditions without significant external disturbances, these parameters exhibit relatively stable correlations. For example, when water flow is stable and the pH remains in the neutral-to-slightly alkaline range, heavy metals tend to exist in a stable state bound to organic matter or sulfides, making them less likely to enter the water. However, disturbances such as heavy rainfall, floods, or pollutant discharge can cause drastic changes in water acidity / alkalinity, electrochemical potential, or flow velocity, disrupting the balance between sediments and the environment, leading to the dissociation of bound heavy metals and their transformation into a soluble state. Therefore, in the initial stage of monitoring, long-term, continuous, and synchronous acquisition of the aforementioned key chemical and physical parameters using high-precision, multi-parameter sensors, and the establishment of a multi-parameter response baseline, can effectively characterize the multi-dimensional coupling relationships under "normal conditions," serving as a basis for judging whether subsequent changes are abnormal and whether there is a risk of pollution. This baseline is comparable, repeatable, and correctable, enabling the system to promptly identify and trigger responses to abnormal states when facing complex natural changes or sudden events, significantly improving the sensitivity and accuracy of the overall sampling system.

[0070] S2, based on a multi-parameter response baseline, acquires monitoring data in real time and performs differential analysis to identify states where the rate of change in pH exceeds a set threshold and is accompanied by significant fluctuations in the proportion of heavy metal bound states, thus determining a period of high risk of potential heavy metal release.

[0071] The steps for identifying high-risk periods for potential heavy metal release from river sediments based on a multi-parameter response baseline include the following sequential operations. This operation involves acquiring monitoring data in real time and comparing it with a previously established multi-parameter response baseline to identify significant changes in the sediment environment from multiple physical and chemical perspectives, thereby determining whether a high-risk state has been entered.

[0072] Real-time monitoring data from multiple sampling profiles are acquired at a set time frequency. Each set of monitoring data includes pH value, electrochemical potential, mass concentration ratios of various heavy metal occurrence forms (including acid-soluble, reducible, oxidizable, and residual forms), water flow velocity, water depth, water temperature, conductivity, dissolved oxygen concentration, and turbidity. Each sampling data set must include a timestamp and sampling profile location identifier. The data sampling frequency is set to once every 15 minutes, and the data acquisition method is automatic collection by multi-parameter sensors deployed at fixed points. To ensure data accuracy, each set of data must undergo a preprocessing stage after acquisition, including numerical rationality verification, missing value imputation, instrument calibration error correction, temperature compensation processing, and is then stored in the river real-time monitoring data record table.

[0073] Each set of data collected in real time is compared item by item with the previously constructed multi-parameter response baseline. This baseline consists of steady-state data collected continuously for at least two weeks under undisturbed conditions, including the normal value range and rate of change range for each parameter. Taking pH as an example, if the difference between the pH values ​​of the second and first sets of samples is greater than 0.5 units and the direction of change is consistently consistent, the rate of change of pH is considered to exceed the upper limit of the normal baseline. Similarly, if the electrochemical potential decreases by more than 50 mV between two adjacent sampling time points, and the concentration of reducible heavy metals in the sediment increases by more than 12% compared to the steady-state reference, then the data set is considered to have abnormal fluctuations linked to multiple parameters. In addition, changes in water flow velocity also need to be comprehensively judged. When a flow velocity increase of more than 0.15 m / s is detected simultaneously, and the turbidity at the sediment interface increases by more than 30%, the judgment signal for the possibility of disturbance is further strengthened.

[0074] When multiple key parameters exhibit simultaneous and drastic changes, the current time period is marked as a "high-risk state for potential heavy metal release." A high-risk state requires three concurrent conditions: first, the pH changes at a rate exceeding 0.5 units within a short period (15 minutes); second, the proportion of reducible or acid-soluble heavy metals increases by more than 10%; and third, the electrochemical potential decreases by more than 30 millivolts, and the water flow velocity increases significantly. When these conditions are met, a high-risk marker is generated in the data record, along with the names of the parameters that triggered the determination, the specific magnitude of the changes, the spatial sampling location, the corresponding sampling time point, and the sampling depth. This high-risk marker is subsequently used to trigger priority sampling strategies and adjust subsequent sampling frequencies, serving as a crucial basis for determining high-risk pollution events.

[0075] To improve the stability of the judgment and reduce the false identification rate, short-term trend analysis is performed on the risk marker information within a continuous time period. Specifically, a continuous time sliding window (e.g., 30 minutes) is set. Within this time range, if at least two sampling records meeting the above judgment conditions are detected, and the trend of the judgment parameters shows a continuously increasing trend, then this time period is confirmed as a genuine high-risk stage for heavy metal release. This stage will serve as the signal source to trigger a multi-level priority sampling mechanism. After each judgment, the system writes the entire judgment process, sampling parameters, and judgment basis of this event into the data archive database in the form of an event record, forming a traceable monitoring decision-making basis. Each record should include the event identification time period, spatial profile number, historical and current values ​​of key parameters, risk level, and recommended response measures.

[0076] The purpose of this step is to achieve rapid identification and dynamic early warning of the environmental state of sediments during monitoring, especially for accurately assessing the risk of heavy metal release caused by sudden changes in pH. In river ecosystems, most heavy metals in sediments exist in a stable bound state, usually forming complexes or precipitates with organic matter, iron oxides, or sulfides, making direct release into the water difficult. However, this stable state is highly sensitive to water pH, electrochemical potential, and hydrodynamic conditions. Especially when subjected to external disturbances (such as acid rain, industrial wastewater discharge, and flooding), the originally stable bound heavy metals in the sediments are easily transformed into soluble forms and enter the water, causing short-term high-concentration pollution peaks. Because such release events are usually characterized by suddenness, short duration, and spatial locality, traditional fixed-point and timed sampling methods are difficult to capture the time and area of ​​occurrence in a timely manner, often leading to missed detections or misjudgments of pollution events. To address this, this step constructs and calls upon a previously established multi-parameter response baseline, and performs differential analysis after acquiring real-time monitoring data, that is, comparing the current monitoring parameters with the baseline data item by item. When the system detects a significant drop in pH value within a short period (e.g., exceeding 0.5 units per hour), accompanied by a significant increase in the proportion of reducible or acid-soluble heavy metals (e.g., exceeding 10%) and a decrease in electrochemical potential, it determines that the current period is in a potentially high-risk state. This step not only has an identification function but also an early warning triggering function, enabling the initiation of subsequent priority sampling or response strategies in the early stages of pollution, achieving rapid response to pollution events. Through this step, the accuracy and response efficiency of the monitoring system in identifying sudden heavy metal release events can be greatly improved, providing important technical support for water environment risk prevention and control, source tracing analysis, and emergency response.

[0077] S3, after identifying a high-risk period, triggers a priority sampling strategy to control the sampling device to dynamically adjust the sampling spatial position and sampling depth, and collects samples from sediment layers with active heavy metal release and areas with high hydrodynamic activity, ensuring the sample acquisition density and representativeness during peak changes.

[0078] Once a high-risk period for potential heavy metal release from sediments is identified through differential analysis, a priority sampling process is immediately initiated. This involves adjusting the spatial location and sampling depth of the sampling device to perform targeted sample acquisition. The specific steps are as follows:

[0079] While the system identifies high-risk states for heavy metal release, it extracts all judgment data for the corresponding time period, including the type of parameters triggering the warning (such as pH decrease rate, electrochemical potential decrease magnitude, and changes in the proportion of heavy metal bound states), the sampling profile number on which the risk judgment is based, the identification time point, and the corresponding sampling depth range. All high-risk profiles are sorted according to their risk level, with higher priority given to profiles that have more triggering parameters, greater fluctuations, and a higher historical frequency of similar disturbances. The sorting results directly affect the scheduling and time allocation of sampling.

[0080] For the priority sampling profiles in the ranking results, their spatial coordinates are obtained, and the lateral movement path and positioning accuracy of the sampling device are controlled accordingly. Using the river centerline as a reference, the sampling device is deployed horizontally to both banks at equal intervals, with three sampling points each, spaced 1 meter apart. Vertically, based on previous sediment structure surveys, the sampling depths are set at 5 cm, 10 cm, and 20 cm below the water-sediment interface. Specific sampling layers are determined by a depth positioning control unit, using a mechanically pressed sampling head to control the depth and ensure that the vertical sampling error does not exceed 0.5 cm. The spatial and depth positioning of the sampling is completed using this method.

[0081] Initiate sediment sampling. The sampling device uses a negative pressure sampling chamber, with each sample extraction volume set to 120 ml to ensure that subsequent testing for multiple indicators such as heavy metal chemical speciation, water content, and organic matter content can be performed. During sampling, each sample is immediately numbered and labeled, with label information including sampling date, sampling time, spatial coordinate point number, vertical depth value, risk level marking, and the signature of the sampling operator. All samples are placed in sealed polyethylene sampling bags and transported in a 4°C refrigerator for no more than 4 hours to ensure that the original state of the samples is not damaged.

[0082] After the initial sampling, the flow velocity, water depth, flow shear stress, and suspended particle concentration at the sampling profile were monitored in real time. Based on data measured by an electromagnetic flowmeter, if the flow velocity was greater than 0.25 m / s, the water shear stress was greater than 0.3 N / m², and the particle concentration in the same profile increased by more than 20 mg / L within 15 minutes, the current flow was deemed to have significant disturbance intensity and the physical ability to drive sediment-released substances to migrate to the upper water layer. In this case, the sampling range was immediately expanded by adding three new sampling profiles downstream along the main flow axis, with each profile spaced 5 meters apart. Depth sampling points consistent with those of the high-risk profile were placed on each new profile, and the same sampling procedure was continued.

[0083] After all sampling operations are completed, all sampling records, real-time monitoring data, spatial distribution information, and sampling labels are organized and archived to establish a priority sampling execution log for this round. The log includes the execution time of each sampling operation, sampling device number, operator signature, sampling coordinates, sampling depth, a summary of the sampling results, and the high-risk identification criteria that triggered the sampling operation. This log information serves as an important reference for subsequent experimental analysis, risk assessment and correction, and optimization of the next round of sampling strategies, ensuring the traceability, completeness, and reverifiability of the sampling data.

[0084] The purpose of this step is to quickly respond and accurately acquire representative samples after identifying critical periods where there is a potential risk of heavy metal release from sediments. This ensures that the monitoring data accurately reflects the timing, spatial range, and release intensity of pollution events. In actual river environments, heavy metals in sediments typically exist stably in a bound state at specific sedimentary layers. When subjected to external disturbances (such as changes in pH or enhanced hydrodynamics), these heavy metals can easily transform from a stable state to a soluble state within a short period and be rapidly released into the upper water layers by the carrying action of water flow, causing short-term high-concentration pollution peaks. However, this process is often sudden, transient, and spatially limited. If fixed-point or fixed-frequency sampling methods are continued, the actual release process may not be captured due to delayed sampling timing or location deviation, leading to monitoring results that underestimate pollution risk or miss pollution sources. Therefore, once a high-risk state is identified, immediately initiating a priority sampling strategy ensures that during the peak of pollution release, priority coverage is given to areas with the most dramatic changes, particularly the active heavy metal release layers within 5 to 20 centimeters below the sediment-water interface, and areas of hydrodynamic disturbance in the river channel characterized by increased flow velocity and shear force. By dynamically adjusting the horizontal position and vertical depth of the sampling device, implementing multi-point, multi-layer sampling, and increasing sample density, the temporal resolution and spatial accuracy of the monitoring data can be effectively improved, thereby ensuring the representativeness and timeliness of the collected samples. This step not only enhances the response speed of the entire monitoring process to sudden pollution events but also significantly strengthens the quantitative analysis capability of the pollution process, providing a data foundation and decision support for pollution source tracing, risk warning, and emergency management.

[0085] S4 performs a rapid chemical speciation analysis on the collected high-density samples to extract the bound state ratio and soluble concentration data of heavy metals, and compares them with the high-risk judgment results. Based on the difference feedback, the risk judgment threshold for heavy metal release is corrected to realize the system's self-correction in response to emergencies.

[0086] After high-density sample collection, rapid chemical speciation analysis is performed on the sediment samples. The risk assessment threshold is then adjusted based on the analysis results and a comparison with previous risk assessments, thereby achieving a self-correcting process for emergency response. This method specifically includes the following steps:

[0087] All sediment samples collected from high-risk heavy metal release profiles were transported to the laboratory for pretreatment. Each sample was processed within 12 hours of refrigerated transport, first undergoing numbering and visual inspection to confirm the absence of obvious leakage or deterioration. Subsequently, the samples were homogenized in a clean environment using a stainless steel scraper, and 10 grams of wet sample were weighed for chemical speciation analysis. The sequential extraction technique described in the national standard HJ 803-2016 "Continuous Extraction Method for Heavy Metal Speciation in Water" was employed, following a four-step extraction process to extract acid-soluble, reducible, oxidizable, and residual heavy metals. The first step involved centrifuging a 0.11 mol / L acetic acid solution at room temperature for 16 hours to extract the acid-soluble metals. The second step used a 0.5 mol / L hydroxylamine hydrochloride solution at 50°C to extract the reducible metals. The third step used a mixture of 30% hydrogen peroxide and 1 mol / L ammonium acetate to extract the oxidizable metals. The fourth step involved wet digestion of the residue with concentrated nitric acid and hydrofluoric acid, followed by determination of the residual metal content. After each extraction step, the extract was filtered through a 0.45 μm pore size membrane, and quantitative analysis was performed using inductively coupled plasma mass spectrometry (ICP-MS) with a detection limit of 0.01 mg / L.

[0088] For each sample, the concentration results of the four heavy metal speciations were analyzed. The proportion of soluble or easily migratable speciations (i.e., acid-soluble and reducible forms) was extracted, and their percentage of the total heavy metal concentration was calculated. This percentage was then compared with the risk identification data at the same time point. The risk identification data was derived from high-risk classifications obtained during the monitoring period based on pH changes, electrochemical potential fluctuations, and historical baseline differential analysis. For example, if a profile was previously identified as having a "severe release risk," and the system predicted that the proportion of reducible forms would be 35% of the total heavy metal content, but the actual measured value was 48%, then the risk identification result underestimated the degree of heavy metal release. Conversely, if the actual measured value was only 26%, it indicated that the previous risk threshold setting might have been too tight, and the relevant judgment conditions needed to be appropriately relaxed to avoid misjudgment.

[0089] Based on the comparison results above, the risk assessment parameters were revised. To avoid frequent fluctuations that could lead to model instability, a fixed-ratio adjustment method was adopted. Specifically, if, within the same monitoring period, the measured values ​​of more than 60% of the samples at "high-risk" assessment points are more than 15% higher than the model's predicted values, a downward adjustment of the risk parameter threshold will be triggered. For example, the original threshold for pH decrease rate was 0.5 units per hour, which was revised to 0.4 units per hour; if the original threshold for the proportion of reducible heavy metals was an increase of 10%, it was adjusted to an increase of 8%. All revisions were recorded in tabular form, including the name of the revised parameter, the original threshold, the revised threshold, the sample number, the revision time, and the basis for the revision, forming a complete risk threshold revision file. This file will be stored in the database as a basis for subsequent auditing and traceability.

[0090] After all judgment parameters have been corrected, the corrected parameters are immediately applied to the judgment process of the next round of real-time monitoring data, forming a feedback loop. The new threshold is used to judge the parameter data collected every 15 minutes in subsequent monitoring cycles to identify in real time whether a high-risk state reappears. If the new judgment result is closer to the proportion of soluble heavy metals in the actual sample, it indicates that the current risk identification model has been optimized; if the deviation is still large, the above feedback correction process continues until the consistency between the judgment result and the measured data reaches the set standard (e.g., the deviation of more than 90% of samples is within ±10%). Through continuous feedback correction, the risk identification mechanism always conforms to the actual environmental characteristics and pollution release process, improving the system's adaptability to different disturbance scenarios.

[0091] The purpose of this step is to rapidly analyze the chemical speciation of collected high-density sediment samples to obtain the proportion of heavy metals in different occurrence forms, especially the true content of soluble and bound heavy metals. These sample analysis results are then compared one by one with the high-risk assessment results made in the previous stage based on a multi-parameter monitoring model. This verifies the accuracy of the risk assessment and allows for timely correction of key risk assessment thresholds when deviations are found, achieving self-correction and dynamic optimization of the monitoring system. The release process of heavy metals in river sediments is complex, influenced by multiple factors such as pH, electrochemical potential, and hydrodynamics, making accurate judgment difficult using a single parameter. Therefore, after priority sampling, continuous extraction and quantitative detection of heavy metals in sediment samples in acid-soluble, reducible, oxidizable, and residual forms can clarify the true release state of pollutants. Comparing these measured data with the risk assessment results made in the previous stage based on parameter mutations (such as rapid pH decrease, potential changes, and abnormal fluctuations in the proportion of bound forms) can identify whether the system model underestimates or overestimates the risk. When the comparison results show that the system's threshold is too lenient or too strict, the triggering conditions of the parameters can be adjusted accordingly. This includes correcting the set values ​​for the pH change rate, electrochemical potential fluctuation amplitude, or heavy metal proportion change amplitude, enabling the system to more accurately and efficiently identify pollution release risks in the next monitoring cycle. This closed-loop feedback mechanism based on sample measurement results gives the entire system the ability to adaptively learn and optimize in real time. It no longer relies on static empirical parameters, significantly improving the sensitivity and reliability of capturing and identifying heavy metal release events under complex disturbances, and ensuring the scientific rigor, accuracy, and on-site adaptability of the entire sampling strategy.

[0092] S5, based on the revised risk assessment threshold, and according to the hydrodynamic and chemical response characteristics of different monitoring areas in the current watershed, the sampling frequency is reallocated, and high-frequency sampling tasks are prioritized for areas with similar disturbance characteristics, so as to achieve early coverage and continuous monitoring of pollution migration paths.

[0093] After revising the threshold for heavy metal release risk assessment, the sampling frequency of each monitoring area was further redistributed based on the hydrodynamic and chemical response characteristics of different monitoring areas within the watershed. This resulted in targeted coverage of areas with similar pollution disturbance trends. The process mainly includes the following steps, aiming to improve overall monitoring efficiency and spatial distribution response capabilities, and strengthen early warning and dynamic tracking of pollution migration processes.

[0094] Hydrodynamic and chemical response data were acquired from all monitoring profiles within the current watershed. Hydrodynamic characteristics included parameters such as flow velocity (in meters per second), flow direction, rate of water level change, and transverse shear force, collected every 15 minutes. Chemical response characteristics included indicators such as pH value, electrochemical potential, and changes in the ratio of reducible to acid-soluble heavy metals, derived from previous sampling analysis and real-time monitoring results. Based on this, a set of numerical indicators describing pollution response sensitivity was constructed for each monitoring point, for example, using the Contaminant Sensitivity Index (CSI) for quantitative calculation.

[0095]

[0096] Where ΔpH represents the change in pH from the baseline, in pH units; ΔE h ΔC represents the change in electrochemical potential, in millivolts. mob The CSI value represents the change in concentration of migratory heavy metals, expressed in milligrams per kilogram; v is the current flow rate (meters per second), and t is the sampling period (minutes). A higher CSI value indicates that the area is more sensitive to pollution disturbances and is more likely to experience heavy metal migration events in subsequent stages.

[0097] All monitoring profiles were grouped and classified according to their CSI values ​​into three categories: high-sensitivity areas, medium-sensitivity areas, and low-sensitivity areas, with corresponding sampling frequency allocation strategies. A basic sampling cycle of 30 minutes was used as the standard; medium-sensitivity areas maintained this frequency, high-sensitivity areas had their sampling cycle reduced to 15 minutes, and low-sensitivity areas had it extended to 60 minutes. The principle for adjusting the sampling frequency was to ensure priority monitoring of high-risk migration paths under the same sampling resources. For example, if the CSI value of a certain section was continuously greater than 1.5 for the past two sampling cycles, and the pH value and the proportion of reducible metals fluctuated drastically, this section would be classified as a high-sensitivity area, and the sampling frequency would be immediately increased. The relevant sampling devices would maintain high-frequency operation for the next two hours, during which a maximum of eight sampling operations could be performed. Conversely, if the river flow velocity in a certain section was less than 0.05 m / s, the electrochemical potential fluctuation was less than 20 mV, and the heavy metal binding state was stable, this section would be classified as a low-sensitivity area, and the sampling frequency could be reduced to minimize resource input and duplicate data.

[0098] The reassigned sampling frequency scheme was applied to the whole-basin monitoring plan and spatially matched with the pollution migration simulation analysis results to confirm whether each high-frequency sampling profile covered the potential pollution propagation path. The downstream impact range L (in meters) of the pollution migration path was approximately calculated based on the measured water flow velocity and direction using the following formula:

[0099] L=f×Δt

[0100] Where f is the average flow velocity, and Δt is the time interval from the release of a risk event to the next sampling. For example, if the flow velocity is 0.3 meters per second when a high-risk event is released, and the subsequent sampling interval is set to 20 minutes, then the possible migration range of the contaminant is:

[0101] L = 0.3 × 20 × 60 = 360 meters

[0102] At this point, high-frequency sampling points should be deployed within a 360-meter downstream range upstream of the event site to ensure coverage of the entire potential pollution migration zone and avoid missed detections or delayed responses to pollution peaks. After the sampling frequency allocation is adjusted, the CSI value and pollution diffusion trend will be reassessed at the end of each complete sampling cycle, and the sampling task plan will be updated in real time based on the latest assessment results.

[0103] The purpose of this step is to dynamically optimize the spatial distribution and temporal scheduling of sampling frequencies after correcting the risk assessment threshold, combining the hydrodynamic and chemical response characteristics of different monitoring areas within the watershed. This enables early coverage of potential pollution migration paths and continuous tracking of heavy metal release events. The core objective of this process is to address the problems of fixed sampling frequencies, rigid regional distribution, and inefficient resource allocation in traditional monitoring. In actual river environments, because hydrodynamic conditions (such as flow velocity, flow direction, and shear force) are highly correlated with pollutant release behavior in sediments, different monitoring points exhibit significant differences in their sensitivity to pollution disturbances. For example, areas with higher flow velocities and flow directions aligned with the downstream of the pollution source are more prone to pollutant migration, while areas with abnormal fluctuations in pH, decreased electrochemical potential, and increased proportions of migrateable heavy metals are more likely to become the main channels for pollutant release and diffusion. By collecting and analyzing these characteristic data, a pollution sensitivity index system can be constructed for each monitoring point, scientifically classifying high-sensitivity, medium-sensitivity, and low-sensitivity areas. Based on this, a zone-specific sampling frequency scheduling system is implemented. Sampling frequencies are increased in highly sensitive areas, such as reducing the original 30-minute interval to 15 minutes, to obtain more intensive monitoring data and achieve early detection and timely tracking of pollution incidents. Sampling frequencies are appropriately reduced in less sensitive areas to avoid wasting resources. This approach not only improves monitoring coverage of key areas but also concentrates sampling resources on areas with the highest risk of pollution migration, effectively enhancing the overall monitoring system's response capabilities, early warning capabilities, and data utilization efficiency. This ensures that sudden heavy metal release events can be accurately identified in their early stages and form a continuous data chain during their propagation, providing scientific support for the entire process of pollution source tracing, diffusion modeling, and emergency decision-making.

[0104] S6 inputs monitoring data and risk response records from the entire process into the adaptive optimization module, updates the multi-parameter response baseline and differential identification model, improves the identification speed and spatial positioning accuracy of pH change events in subsequent operation, and builds a closed-loop response mechanism with dynamic control capabilities.

[0105] To improve the monitoring accuracy and response efficiency of heavy metal release events from river sediments, a monitoring process with closed-loop response capability is achieved by incorporating monitoring data and risk response records throughout the entire process into a continuously updated workflow. This involves dynamically adjusting the previously established multi-parameter response baseline and abrupt change identification rules. The specific steps of this implementation are as follows:

[0106] All raw data and processing records generated during the current monitoring cycle are collected and summarized. The data includes: pH value, electrochemical potential, flow rate, water level, temperature, dissolved oxygen, and conductivity for each sampling profile; and the mass concentration (in milligrams per kilogram) of heavy metals in different occurrence forms (acid-soluble, reducible, oxidizable, and residual) obtained from sediment sample analysis. Simultaneously, records of every adjustment to the sampling device throughout the cycle are collected, including the time of sampling frequency change, sampling depth adjustment time, criteria for initiating sampling, spatial coordinates of sampling points, and risk identification level for the sampling time period. All data is automatically time-stamped and includes sampling location coordinates during acquisition, and is ultimately archived as a unified format data file and stored in the data record database.

[0107] The collected data is periodically filtered and calculated to update the background values ​​of various environmental parameters in the response baseline. Data from all time periods identified as "non-high-risk" within the past 48 hours are selected, and the maximum, minimum, average, and standard deviation of each physical and chemical parameter are calculated. Taking pH as an example, the normal fluctuation range (e.g., 6.85–7.25) is determined by continuously calculating the pH data from all sampling points within that time period. Statistical analysis of the concentration data of reducible heavy metals is performed to determine their average mass concentration and upper and lower limits (e.g., 30–45 mg / kg) under steady-state conditions. If the background value of the new round shows a significant change compared to the previous period (e.g., the difference between the upper and lower limits exceeds 10%), the background value of this round is used as the new multi-parameter response baseline and recorded in the response baseline update table for use in the next round of identification.

[0108] Based on the updated response baseline, the difference identification criteria are simultaneously revised. The difference identification criteria refer to the reference standards for determining whether a sudden event has occurred, and must consider the rate of parameter change, the magnitude of change, and the interconnectedness of changes. In this step, the mutation judgment criteria are reset based on the upper and lower limits of the new baseline. For example, if the new data shows that the change in electrochemical potential under normal conditions has decreased to within 20 mV, the original "decline of 50 mV" trigger condition will no longer be applicable, and the trigger threshold needs to be lowered to "decline of 30 mV". Similarly, if the stable fluctuation range of reducible heavy metals in sediments changes from 30–45 mg / kg to 25–35 mg / kg, the judgment criterion for a "drastic increase" will be adjusted from "increase of 10 mg / kg" to "increase of 6 mg / kg". All revisions will be recorded in the mutation identification criterion revision record table with textual descriptions and numerical comparisons, replacing the old criterion table, for risk assessment in the next monitoring cycle.

[0109] The updated response baseline and identification criteria are used as basic inputs and deployed into subsequent monitoring processes, forming a closed-loop operation logic of monitoring-judgment-adjustment-remonitoring. All newly collected data is directly compared with the updated baseline after collection. If any parameter shows a trend exceeding the normal range, a risk event is determined according to the new criteria. If determined to be a high-risk event, the current response mechanism immediately initiates processes such as increasing sampling frequency, adjusting spatial distribution, and rapid chemical speciation analysis. After the next monitoring cycle, the full-cycle data is input into the aforementioned process again, continuously forming an updated closed loop, enabling the system to dynamically adapt to and optimize in real time to environmental changes. This mechanism can maintain high sensitivity to pollution emergencies over a long period and improve location accuracy and judgment reliability.

[0110] This step aims to integrate multi-source data collected throughout the monitoring process with corresponding risk response behavior records, enabling continuous updating and optimization of multi-parameter response baselines and identification rules. This constructs a closed-loop response mechanism with adaptive control capabilities. The core value of this mechanism lies in enabling the monitoring system to move away from fixed static thresholds or manually set empirical parameters. Instead, it relies on the accumulation and analysis of measured data to dynamically adjust identification rules, thereby continuously improving the speed and spatial positioning accuracy of judging heavy metal release events caused by sudden changes in pH. In actual river environments, the pH, electrochemical potential, and heavy metal occurrence states in sediments are coupled with hydrodynamic conditions, exhibiting nonlinear, time-varying, and regionally variable changes. If the system uses initially set response baselines and differential thresholds for an extended period, it will inevitably lose accuracy as the environmental context changes. Through this step, the system constructs a complete "data-response-verification" chain by combining data such as pH, electrochemical potential, flow velocity, shear force, and concentrations of reducible and acid-soluble heavy metals in sediments from each sampling cycle, along with information such as the spatial coordinates, judgment indicators, sampling initiation time, and sampling results for each response trigger. The system also periodically updates the response baseline and triggering rules. For example, if it finds that high-risk events have been frequently triggered recently with a pH decrease of 0.4 units instead of 0.5 units, the system will automatically adjust the pH trigger threshold to more sensitively identify potential events. Simultaneously, the system will overlay spatial response results with hydrodynamic data analysis to correct the spatial weight ranking of high-risk points in the river channel, improving positioning accuracy. After multiple rounds of closed-loop operation, the system can gradually establish a risk identification model with regional adaptability and real-time learning capabilities, realizing a "collection-judgment-response-correction-rejudgment" cyclic mechanism. This enhances the stability and accuracy of the monitoring system in complex aquatic environments, providing solid technical support for long-term dynamic monitoring and rapid response to sudden pollution events.

[0111] The adaptive sampling method for multi-parameter sensing of river sediments proposed in this invention enables highly timely, spatially resolved, and representative monitoring of the instantaneous release of heavy metals from sediments. This invention significantly improves the early detection capability of abrupt pH changes by establishing a multi-parameter response baseline at the initial monitoring stage and introducing a linkage identification mechanism involving pH, electrochemical potential, heavy metal binding state ratio, and hydrodynamic characteristics during sampling. Based on this, by dynamically adjusting sampling location and frequency, accurately identifying high-risk layers, and rapidly conducting chemical speciation analysis, combined with self-calibration and threshold update mechanisms, the system ensures rapid response and real-time correction of judgment criteria during critical periods of heavy metal release, effectively avoiding data omissions and misjudgments. Finally, through full-process data feedback and dynamic optimization of the differential identification model, a closed-loop response mechanism with self-learning and adaptive capabilities is constructed, comprehensively improving the accuracy, reliability, and intelligence level of heavy metal pollution monitoring in river sediments, providing solid data support and decision-making basis for pollution early warning, source tracing, and ecological restoration.

[0112] This invention provides, for example Figure 2 The adaptive sampling system for multi-parameter sensing of river sediments shown includes a multi-parameter baseline construction module, a risk identification and differential judgment module, a priority sampling control module, a rapid morphological analysis and threshold correction module, a regional frequency scheduling module, and a closed-loop optimization and model update module.

[0113] The multi-parameter baseline construction module synchronously collects sediment pH, electrochemical potential, heavy metal bound state ratio, water flow velocity and hydrodynamic parameters through sensing devices to establish a multi-parameter response baseline.

[0114] The risk identification and differential judgment module, based on a multi-parameter response baseline, collects monitoring data in real time and performs differential analysis to identify high-risk periods where the rate of change in pH exceeds the threshold and the proportion of heavy metal bound states fluctuates.

[0115] The priority sampling control module identifies high-risk periods and triggers a priority sampling strategy to adjust the spatial position and depth of the sampling device to collect samples from layers with active heavy metal release and high hydrodynamic regions.

[0116] The rapid morphology analysis and threshold correction module performs rapid chemical morphology analysis on the collected samples to obtain the proportion of heavy metal bound states and the concentration of soluble states. Based on the comparison results, it corrects the risk judgment threshold to achieve system self-calibration.

[0117] The regional frequency scheduling module, based on threshold correction, reallocates sampling frequencies according to the hydrodynamic and chemical characteristics of each monitoring area, prioritizing scheduling to areas with similar disturbance characteristics to achieve pollution migration path coverage;

[0118] The closed-loop optimization and model update module inputs the monitoring data and response records of the entire process into the adaptive optimization module to update the multi-parameter response baseline and differential identification model, and build a closed-loop response mechanism with dynamic control capabilities.

[0119] The adaptive sampling method for multi-parameter sensing of river sediments provided in this embodiment of the invention is implemented through the aforementioned adaptive sampling system for multi-parameter sensing of river sediments. For details of the specific methods and procedures of the adaptive sampling system for multi-parameter sensing of river sediments, please refer to the embodiments of the aforementioned adaptive sampling method for multi-parameter sensing of river sediments, which will not be repeated here.

[0120] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An adaptive sampling method for multi-parameter sensing of river channel sediments, characterized in that, Includes the following steps: S1, by synchronously collecting the pH, electrochemical potential, heavy metal binding state ratio of sediments, as well as water flow velocity and hydrodynamic parameters through sensing devices, to establish a multi-parameter response baseline; S2, based on a multi-parameter response baseline, collects monitoring data in real time and performs differential analysis to identify high-risk periods when the rate of change in pH exceeds the threshold and the proportion of heavy metal bound states fluctuates. S3, after identifying high-risk periods, triggers a priority sampling strategy, adjusts the spatial position and depth of the sampling device, and collects samples from layers with active heavy metal release and high hydrodynamic areas. S4 performs rapid chemical speciation analysis on the collected samples to obtain the proportion of heavy metal bound state and the concentration of soluble state. Based on the comparison results, the risk judgment threshold is corrected to achieve system self-calibration. S5, based on threshold correction, redistributes sampling frequencies according to the hydrodynamic and chemical characteristics of each monitoring area, prioritizing areas with similar disturbance characteristics to achieve pollution migration path coverage; S6 inputs the full-process monitoring data and response records into the adaptive optimization module, updates the multi-parameter response baseline and differential identification model, and constructs a closed-loop response mechanism with dynamic control capabilities.

2. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 1, characterized in that, Step S1 includes: Multiple monitoring points were set up at representative locations in the target monitoring section of the river channel. Each monitoring point was equipped with multiple cross-sectional and vertical sampling points, and pH sensors, electrochemical potential electrodes, conductivity sensors, flow velocity sensors and water depth sensors were installed. The proportions of acid-soluble, reducible, oxidizable, and residual heavy metals in sediments were obtained by using a water sample extraction device combined with sequential extraction and quantitative analysis. The data collection frequency was set to once every 15 minutes, with a continuous collection period of no less than 14 days, and the collected data covered typical hydrological processes. Parametric analysis is performed on structured sample data to extract the mean, standard deviation and variation under natural steady-state conditions, establish a response baseline with clear numerical boundaries, and revise and update it monthly.

3. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 1, characterized in that, Step S2 includes: Monitoring data including pH, electrochemical potential, mass concentration ratio of various heavy metal forms, water flow velocity, water depth, water temperature, conductivity, dissolved oxygen concentration and turbidity are acquired at a time frequency of 15 minutes. The collected data is preprocessed and compared item by item with the multi-parameter response baseline; When the rate of change in pH exceeds 0.5 units, the electrochemical potential decreases by more than 30 millivolts, the proportion of heavy metals in reducible or acid-soluble form increases by more than 10%, and the water flow velocity increases significantly, it is identified as a potentially high-risk period. Sliding window trend analysis is performed on high-risk states over a continuous period of time to identify high-risk phases and record the basis for the judgment.

4. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 3, characterized in that, The specific steps for performing sliding window trend analysis on high-risk states over a continuous time period to identify high-risk phases and record the criteria for judgment are as follows: Set a sliding analysis window with a time length of 30 minutes to perform a coverage scan of continuously sampled data; Within each sliding window, record at least two instances of the combined criteria of pH change rate, electrochemical potential decrease, heavy metal occurrence speciation change, and water flow enhancement are collected. When the judgment parameters show a continuous increasing trend within the window time, it is confirmed as a high-risk stage of real heavy metal release, and the time interval of risk judgment, sampling profile number, key parameter change value and recommended response measures are recorded.

5. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 1, characterized in that, Step S3 includes: Based on the high-risk assessment results, extract the trigger parameter type, sampling profile number, identification time point, and corresponding sampling depth; High-risk profiles were sorted according to risk level, and multiple sampling points were set up laterally along the river channel with a fixed vertical sampling depth. Sediment samples were extracted using a negative pressure sampling chamber and labeled. The samples were transported in a 4°C refrigerated environment. When the flow rate and particulate matter concentration reach the disturbance conditions, the sampling profile is expanded along the downstream main axis and the sampling operation is performed simultaneously. After sampling is completed, a sampling execution log containing time, coordinates, depth, and triggering basis is created and archived.

6. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 1, characterized in that, Step S4 includes: Sediment samples collected from high-risk profiles were numbered, checked, and pretreated. Acid-soluble, reducible, oxidizable, and residual heavy metals were extracted using a sequential extraction method, and their concentrations were determined using inductively coupled plasma mass spectrometry. Extract the proportion of acid-soluble and reducible metal concentrations and compare it with the risk identification and judgment results; Based on the comparison deviation, the risk assessment threshold is adjusted by a fixed percentage to form a risk threshold correction file; The revised parameters are applied to the next round of real-time monitoring to determine high-risk states, forming a closed-loop feedback process.

7. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 1, characterized in that, Step S5 includes: Collect hydrodynamic and chemical response parameters from all monitoring profile points and calculate the pollution sensitivity index (CSI) value. Based on the CSI value, the monitoring area is divided into high-sensitivity zone, medium-sensitivity zone and low-sensitivity zone, and sampling cycles of 15 minutes, 30 minutes and 60 minutes are set respectively. The actual migration distance of the pollution downstream is calculated based on the current flow rate and sampling interval. High-frequency sampling points are then deployed along the main flow direction within this distance range to ensure full coverage and continuous tracking of the pollution migration path.

8. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 1, characterized in that, Step S6 includes: The data on pH, electrochemical potential, flow rate, water level, temperature, dissolved oxygen, conductivity, and concentration of various forms of heavy metals during the current monitoring period, along with the sampling and adjustment records, will be archived together. The maximum, minimum, average, and standard deviation of non-high-risk status data within 48 hours are calculated to form a new multi-parameter response baseline. Based on the new baseline adjustment difference identification criteria, the trigger thresholds of each judgment parameter are numerically corrected and an identification criterion correction record table is formed; The updated multi-parameter response baseline and identification criteria are used as the basis for judgment and input into the next round of monitoring process to build a dynamically updated closed-loop response mechanism.

9. The adaptive sampling method for multi-parameter sensing of river sediments according to claim 8, characterized in that, The specific steps for numerically correcting the trigger thresholds of each judgment parameter and forming a correction record table for the identification criteria based on the new baseline adjustment difference identification criteria are as follows: Statistical analysis was performed on the variation range of each monitoring parameter within the normal fluctuation range. The trigger thresholds for pH value, electrochemical potential, and reducible heavy metal concentration were reset. The parameter names before and after the correction, the original threshold, the corrected threshold, the correction time, and the data on which they were based were recorded in the identification criterion correction record table. This record was used to replace the original judgment criteria and guide the next round of risk identification and judgment process.

10. An adaptive sampling system for multi-parameter sensing of river sediments, used to achieve the features described in any one of claims 1-9 above. An adaptive sampling method for multi-parameter sensing of channel sediments is characterized by comprising a multi-parameter baseline construction module, a risk identification and differential judgment module, a priority sampling control module, a rapid morphological analysis and threshold correction module, a regional frequency scheduling module, and a closed-loop optimization and model update module; wherein, The multi-parameter baseline construction module synchronously collects sediment pH, electrochemical potential, heavy metal bound state ratio, water flow velocity and hydrodynamic parameters through sensing devices to establish a multi-parameter response baseline. The risk identification and differential judgment module, based on a multi-parameter response baseline, collects monitoring data in real time and performs differential analysis to identify high-risk periods where the rate of change in pH exceeds the threshold and the proportion of heavy metal bound states fluctuates. The priority sampling control module identifies high-risk periods and triggers a priority sampling strategy to adjust the spatial position and depth of the sampling device to collect samples from layers with active heavy metal release and high hydrodynamic regions. The rapid morphology analysis and threshold correction module performs rapid chemical morphology analysis on the collected samples to obtain the proportion of heavy metal bound states and the concentration of soluble states. Based on the comparison results, it corrects the risk judgment threshold to achieve system self-calibration. The regional frequency scheduling module, based on threshold correction, reallocates sampling frequencies according to the hydrodynamic and chemical characteristics of each monitoring area, prioritizing scheduling to areas with similar disturbance characteristics to achieve pollution migration path coverage; The closed-loop optimization and model update module inputs the monitoring data and response records of the entire process into the adaptive optimization module to update the multi-parameter response baseline and differential identification model, and build a closed-loop response mechanism with dynamic control capabilities.

Images