Intelligent saline-alkali soil monitoring sensor drift self-calibration system and method

Abstract

The application discloses a kind of intelligent saline-alkali soil monitoring sensor drift self-calibration system and method;The system includes multi-parameter monitoring node, redundancy comparison module, miniature soil water sample collection and comparison unit, environmental coupling drift identification module, calibration compensation algorithm module and remote communication and management platform;The present application generates the first drift identification result by main / sensor cross comparison, obtains reference datum by extracting soil water sample on site, and generates the second drift identification result by comparing the expected response value with the measured value calculated by the environmental coupling response model;Calibration compensation algorithm module calculates calibration coefficient by the above results, temperature compensation, environmental compensation and proportion calibration are carried out to original measurement value, and the final calibration value is output;Remote platform supports OTA parameter update;The present application can automatically identify and compensate sensor drift, improve the accuracy and reliability of long-term monitoring data in high-salt and high-corrosion environment of saline-alkali soil, and reduce operation and maintenance cost.

Description

Technical Field

[0001] This invention belongs to the fields of soil monitoring, environmental sensing and intelligent agricultural safety technology, and specifically relates to an intelligent saline-alkali land monitoring sensor drift self-calibration system and method. Background Technology

[0002] Saline-alkali land is widely distributed in arid and semi-arid regions, and its soil salinity and alkalinity exhibit significant spatiotemporal variations. Long-term monitoring of saline-alkali land is crucial for agricultural production, saline-alkali land improvement, groundwater management, and ecological restoration. Existing monitoring systems commonly use buried sensors, such as electrical conductivity (EC) sensors, pH electrodes, ion-selective electrodes (ISE), and soil moisture probes. However, the typical environmental characteristics of saline-alkali land, including high salinity, high alkalinity, and high evaporation, pose significant challenges to long-term sensor monitoring.

[0003] Under high evaporation conditions, salts easily crystallize and deposit on the probe surface, causing measurement drift; simultaneously, the Na+ in saline-alkali soil... + CO3² - HCO3 - Cl - Plasma accelerates electrode corrosion, gradually increasing long-term monitoring errors. This is especially true for Na. + / Ca² + Plasma-selective electrodes exhibit rapid membrane potential decay and significant drift in high-salt environments. Furthermore, the complex coupling between temperature, water content, and salinity leads to substantial differences in the response curves of the same sensor under varying environmental conditions, further complicating data interpretation.

[0004] Currently, most existing monitoring equipment relies on manual periodic calibration, requiring professionals to bring standard solutions to the site to disassemble, clean, and calibrate the sensors. This approach is not only labor-intensive and inefficient, but also results in unreliable data during calibration intervals. More importantly, existing systems lack the ability to automatically identify and calibrate sensor drift online, failing to distinguish between data fluctuations caused by environmental changes and drift due to sensor performance degradation, leading to misjudgments or missed detections. Therefore, there is an urgent need for an intelligent monitoring system capable of long-term unattended operation with autonomous drift identification and online calibration functions. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent self-calibration system and method for monitoring the drift of saline-alkali land sensors.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This application provides an intelligent self-calibration system for drift of a saline-alkali land monitoring sensor, comprising: A multi-parameter monitoring node is used to collect various soil parameters at the monitoring location, including at least electrical conductivity, pH value, ion concentration, temperature and moisture content. A redundancy comparison module, connected to the multi-parameter monitoring node, is used to cross-compare the output values ​​of the main sensor and the auxiliary sensor at the same monitoring location, and generate a first drift identification signal based on the comparison result. A miniature soil water sample collection and comparison unit includes a sampling pump, a filter, a flow-through comparison chamber, and a reference electrode disposed in the comparison chamber. This unit is used to extract water samples from the soil and obtain reference reference values ​​through the reference electrode in the flow-through comparison chamber. An environmental coupling drift identification module is used to acquire environmental variables, calculate the expected response value using an environmental coupling response model, and compare the actual measured value of the sensor with the expected response value to generate a second drift identification signal. The calibration compensation algorithm module is used to determine the calibration coefficient based on the first drift identification signal, the second drift identification signal and the reference value, and to perform temperature compensation, environmental compensation and proportional correction on the original sensor measurement value, and output the final calibration value. The remote communication and management platform communicates with the above modules and is used to receive the final calibration value and update the system operating parameters or processing logic remotely.

[0007] Preferably, the multi-parameter monitoring node includes: a conductivity sensor forming a primary-secondary pairing relationship, a pH sensor forming a primary-secondary pairing relationship, at least one ion-selective electrode, and a temperature / moisture content sensor; and the paired sensors are different from each other in terms of range, structure, or measurement mechanism.

[0008] Preferably, the micro soil water sample collection and comparison unit further includes a standard solution storage and injection component, used to periodically inject standard solution into the flow-through comparison chamber to achieve online calibration of the reference electrode.

[0009] Preferably, the environmental coupling response model uses at least one of temperature, soil moisture content, salinity, and ionic strength as input parameters, and calculates the expected response value through a multinomial regression model, a random forest model, or a neural network model.

[0010] Preferably, the calibration compensation algorithm module is further used to calculate the health index of the sensor, wherein the health index H is determined according to the following formula: in, For the first The deviation between the sensor measurement value and the reference value or the expected response value of the model in the comparison. Specify the sensor's rated range or a reference value. The number of comparisons included in the statistics; the calibration compensation algorithm module in the health index When the value falls below a set threshold, a maintenance alarm is issued through the remote communication and management platform.

[0011] A sensor drift self-calibration method for monitoring saline-alkali land includes the following steps: Collect real-time sensor measurements and environmental variables output from multi-parameter monitoring nodes; The measurements from the main sensor and the auxiliary sensor at the same monitoring location are compared. When the difference between the two exceeds a first threshold, a first drift identification result is generated. The sampling pump is started to extract soil water samples, which are then filtered and introduced into the flow-through comparison chamber. The reference value is measured using the reference electrode located in the comparison chamber. The actual measured value of the sensor is compared with the expected response value calculated by the environmental coupling response model based on the environmental variables. When the deviation exceeds the second threshold, a second drift identification result is generated. Based on the reference value and the first and second drift identification results, the calibration coefficient is calculated, and temperature compensation and environmental compensation are applied to the original measurement values ​​respectively. The final calibration value is calculated and output based on the calibration coefficient, temperature compensation, and environmental compensation. The final calibration value is uploaded to the remote communication and management platform, and the local operating parameters are updated according to the platform's instructions.

[0012] Preferably, the calibration coefficient K is calculated according to the following formula: in, The reference base value, This represents the measurement value of the sensor to be calibrated at the sampling time.

[0013] Preferably, the final calibration value Calculate according to the following formula: in, The original measurement value of the sensor. This is the temperature compensation amount. For environmental compensation, This is the calibration coefficient.

[0014] Preferably, the acquisition period of the reference value is dynamically adjusted according to the sensor's health index or historical drift trend; the health index is calculated as follows: in, For the first The deviation between the sensor measurement value and the reference value or the expected response value of the model in the comparison. For the rated range or specified reference value, This represents the number of comparisons.

[0015] Preferably, the remote update performed by the remote communication and management platform includes at least one of the following: updating the parameters of the environmental coupling response model, modifying the first threshold or the second threshold, adjusting the data acquisition frequency of the sensor, or issuing updated values ​​of the calibration coefficients.

[0016] Compared with the prior art, this application has the following beneficial effects: This invention proposes an intelligent self-calibration system and method for monitoring sensor drift in saline-alkali land. By constructing a system architecture that includes multi-parameter monitoring nodes, a redundant comparison module, a micro soil and water sample collection and comparison unit, an environmental coupling drift identification module, a calibration compensation algorithm module, and a remote communication and management platform, it can automatically identify measurement drift caused by sensor aging, corrosion, salt crystallization, etc., and distinguish it from normal environmental fluctuations. It uses soil and water samples extracted on-site as a benchmark to calculate online calibration coefficients, and dynamically compensates for measured values ​​by combining temperature and environmental coupling models. At the same time, it achieves high-reliability monitoring under long-term unattended operation through health index assessment and remote OTA updates, thereby significantly improving the long-term stability and accuracy of sensor data in the high-salt and high-corrosion environment of saline-alkali land, and greatly reducing the frequency of manual calibration and operation and maintenance costs. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the system of the present invention; Figure 2 Schematic diagram of a redundant monitoring unit with main / auxiliary dual probes; Figure 3 This is a schematic diagram of a micro water sample collection and benchmark comparison unit; Figure 4 Flowchart for environmental coupling drift identification; Figure 5 Flowchart for calibration compensation algorithm processing; Figure 6 Flowchart for sensor health index assessment and alarm; Figure 7 This is a hardware structure diagram of a multi-parameter monitoring node. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Furthermore, in this invention, an element referred to as fixed to or disposed on another element may be directly disposed on the other element, or there may be an intermediate element; when an element is considered to be connected to another element, it may be directly connected to the other element, or there may be an intermediate element at the same time; the terms vertical, horizontal, left, right and similar expressions used herein are for illustrative purposes only and do not represent the only embodiments. See Figures 1-7 This application provides an intelligent self-calibration system for drift of a saline-alkali land monitoring sensor, comprising: A multi-parameter monitoring node is used to collect various soil parameters at the monitoring location, including at least electrical conductivity, pH value, ion concentration, temperature and moisture content. A redundancy comparison module, connected to the multi-parameter monitoring node, is used to cross-compare the output values ​​of the main sensor and the auxiliary sensor at the same monitoring location, and generate a first drift identification signal based on the comparison result. A miniature soil water sample collection and comparison unit includes a sampling pump, a filter, a flow-through comparison chamber, and a reference electrode disposed in the comparison chamber. This unit is used to extract water samples from the soil and obtain reference reference values ​​through the reference electrode in the flow-through comparison chamber. An environmental coupling drift identification module is used to acquire environmental variables, calculate the expected response value using an environmental coupling response model, and compare the actual measured value of the sensor with the expected response value to generate a second drift identification signal. The calibration compensation algorithm module is used to determine the calibration coefficient based on the first drift identification signal, the second drift identification signal and the reference value, and to perform temperature compensation, environmental compensation and proportional correction on the original sensor measurement value, and output the final calibration value. The remote communication and management platform communicates with the above modules and is used to receive the final calibration value and update the system operating parameters or processing logic remotely.

[0019] In this embodiment, the multi-parameter monitoring node is specifically implemented as follows: the multi-parameter monitoring node is buried at different depths (e.g., 10cm, 30cm, 50cm) in saline-alkali soil to collect various soil parameters at the monitoring location in real time; the various soil parameters include at least electrical conductivity (EC), pH value, and ion concentration (e.g., Na+). + Ca²+ The node integrates a primary / secondary EC sensor, a primary / secondary pH sensor, at least one ion-selective electrode, and a temperature / moisture content sensor. Each sensor is connected to a data acquisition and processing unit (MCU), which is responsible for analog signal acquisition, preprocessing, storage, and preliminary diagnosis. The node is also equipped with a power management unit that can be connected to a solar panel or battery to support long-term low-power operation. The node collects data at a set period (e.g., every 15 minutes).

[0020] The redundancy comparison module is implemented as follows: Connected to a multi-parameter monitoring node, it performs cross-comparison of the output values ​​of the main sensor and auxiliary sensor at the same monitoring location. This module calculates the output E of the main sensor. A With auxiliary sensor output E B The absolute value of the difference ΔE = |E A -E B When ΔE exceeds the preset first threshold θ1, a first drift identification signal is generated. The threshold θ1 can be adaptively set according to the sensor range, resolution and historical statistical fluctuation range, or it can be dynamically adjusted by the remote management platform. This module can be calculated by the MCU inside the node, or it can be implemented on the cloud platform for group statistical comparison.

[0021] The specific implementation of the miniature soil water sampling and comparison unit: This unit includes a sampling pump, a filter, a flow-through comparison chamber, and a reference electrode set in the comparison chamber; the sampling pump is a miniature water pump used to extract a small amount of soil water sample (e.g., 1-10 mL) from a preset depth or monitoring well; the filter is used to remove larger particles and suspended solids from the water sample to prevent clogging of the reference electrode; the inlet of the flow-through comparison chamber is connected to the outlet of the filter, and at least one set of high-precision reference electrodes (e.g., laboratory-grade EC / pH electrodes or periodically replaced fresh electrodes) is installed in the chamber; the system controls the operation of the sampling pump to draw the water sample into the comparison chamber, and the reference value E is measured by the reference electrode. ref The water sample is then drained or returned to the soil. The unit can also be equipped with a reference solution injection device to periodically inject standard solution into the comparison chamber to achieve automatic calibration of the reference electrode itself.

[0022] The specific implementation of the environmental coupling drift identification module: This module acquires environmental variables (including temperature T, water content W, salinity S, ionic strength I, etc.) and calculates the expected response value E based on the environmental coupling response model. true The response model can be a multinomial regression model, a random forest model, or a neural network model, with inputs T, W, S, and I, and output E. true =f(T,W,S,I); the module will convert the actual sensor measurement value E meas With E trueCompare, when E is satisfied true -E meas When the threshold θ2 is greater than θ2, a second drift identification signal is generated; the threshold θ2 can be dynamically adjusted based on historical data and error statistics, and the model parameters can be obtained through offline training using laboratory calibration data or field operation data.

[0023] The specific implementation of the calibration compensation algorithm module: This module is based on the first drift identification signal, the second drift identification signal, and the reference value E. ref Calculate the calibration coefficient K, and perform temperature compensation, environmental compensation, and proportional calibration on the original sensor measurements; specifically: The calibration factor K is based on K=E ref / E meas Calculate and store the values ​​in the calibration parameter table; subsequent measurements are then proportionally corrected according to K. Temperature compensation: for the original measured value E raw Subtract the temperature compensation amount ΔE temp (T), to obtain E temp_corr =E raw -ΔE temp (T), where ΔE temp (T) is a linear or nonlinear temperature compensation function; Environmental compensation: E output from the environmental coupling model true With E temp_corr The difference is used to calculate the environmental compensation amount ΔE. env E env_corr =E temp_corr +ΔE env ; Final calibration output value E out =K·(E raw -ΔE temp )+ΔE env .

[0024] The specific implementation of the remote communication and management platform: This platform communicates with the above modules (using LoRa, NB-IoT, or 4G communication methods) to receive calibrated data; the platform is deployed on a local server or cloud server to realize functions such as historical data storage, visualization, threshold alarms, and sensor health status management; at the same time, the platform supports the distribution of updated system operating parameters or processing logic to the front-end nodes via OTA (Over-The-Air), including updating response model parameters, correcting thresholds, adjusting the acquisition frequency, or distributing new calibration coefficient upper limits.

[0025] In a preferred embodiment, the multi-parameter monitoring node includes: a conductivity sensor forming a primary-secondary pairing relationship, a pH sensor forming a primary-secondary pairing relationship, at least one ion-selective electrode, and a temperature / moisture content sensor; and the paired sensors are different from each other in terms of range, structure, or measurement mechanism. In this embodiment, the main conductivity sensor uses an electrode-type probe, and the auxiliary conductivity sensor uses an inductive probe; the two have different measurement ranges and different measurement mechanisms. The main pH sensor and the auxiliary pH sensor can use different materials (e.g., glass electrodes and solid-state electrodes) or have different structures. An ion-selective electrode is used to measure Na+. + Ca² + HCO3 - or Cl - One or more of the following; temperature / moisture content sensors are used to monitor changes in the soil environmental field; because the main / auxiliary sensors differ in range, structure, materials or measurement mechanism, their response characteristics to the same environmental conditions are different. When one of them drifts, the anomaly can be quickly identified by calculating the difference ΔE, which improves the reliability and accuracy of drift identification.

[0026] In a preferred embodiment, the micro soil water sample collection and comparison unit further includes a standard solution storage and injection component, which is used to periodically inject standard solution into the flow-through comparison chamber to achieve online calibration of the reference electrode.

[0027] In this embodiment, an optional automatic reference solution injection module is included as part of the unit, which incorporates a reference solution storage tank and an injection unit. When needed (e.g., at regular intervals or when an abnormal response of the reference electrode is detected), the system controls the injection device to inject a standard solution of known concentration into the flow-through comparison chamber. The reference electrode measures the standard solution and compares it with its theoretical value, thereby calculating the drift of the reference electrode itself and performing internal corrections. This process ensures the accuracy of the reference electrode during long-term use, thus ensuring the accuracy of the reference value E. ref Reliability.

[0028] In a preferred embodiment, the environmental coupled response model uses at least one of temperature, soil moisture content, salinity, and ionic strength as input parameters, and calculates the desired response value through a multinomial regression model, a random forest model, or a neural network model.

[0029] In this embodiment, the input variables of the environmental coupling response model include temperature T, water content W, salinity S, and ionic strength I; the model form can be a multinomial regression model (e.g., It can also be a random forest model or a neural network model; the model is trained offline using laboratory calibration data or field operation data, and after training, it is deployed on the front-end device. During runtime, the module quickly calculates the expected response value based on real-time collected environmental variables. This model can capture the impact of multi-factor coupling on sensor output in saline-alkali environments, thus effectively distinguishing between normal environmental fluctuations and sensor performance degradation.

[0030] In a preferred embodiment, the calibration compensation algorithm module is further used to calculate the health index of the sensor, wherein the health index H is determined according to the following formula: in, For the first The deviation between the sensor measurement value and the reference value or the expected response value of the model in the comparison. Specify the sensor's rated range or a reference value. The number of comparisons included in the statistics; the calibration compensation algorithm module in the health index When the value falls below a set threshold, a maintenance alarm is issued through the remote communication and management platform.

[0031] In this embodiment, the health index H is calculated as follows: Assume a total of n comparisons were performed, and the deviation between the sensor measurement value and the reference value (or the model's expected value) in each comparison is... , This refers to the sensor's rated range or a specified reference value (e.g., the sensor's full-scale value); then When H is below a preset threshold (e.g., 0.7), the system determines that the sensor has entered a fault or critical state and automatically sends an alarm message to the remote communication and management platform, prompting the management personnel to arrange on-site maintenance or replacement of the sensor. The health index is not only used for alarms, but can also be used to dynamically adjust the calibration frequency based on its value, thereby extending the effective service life of the sensor.

[0032] On the other hand, this application provides a sensor drift self-calibration method for monitoring saline-alkali land, including the following steps S1-S7: Step S1. Collect real-time sensor measurements and environmental variables output by the multi-parameter monitoring nodes.

[0033] Similar to the system implementation method, the multi-parameter monitoring node collects data such as conductivity, pH, ion concentration, temperature, and water content at a preset cycle (e.g., 15 minutes). Among these, temperature, water content, salinity, and ion strength are also used as environmental variables for backup.

[0034] Step S2. Compare the measurement values ​​of the main sensor and the auxiliary sensor at the same monitoring location. When the difference between the two exceeds the first threshold, the first drift identification result is generated.

[0035] Compare the measurements from the main sensor and the auxiliary sensor at the same monitoring location, and calculate the difference ΔE = |E A -E BWhen ΔE exceeds the first threshold θ1, a first drift identification result is generated; the threshold θ1 can be set as a percentage of the sensor range (e.g., 5%), or it can be dynamically adjusted by the platform.

[0036] Step S3. Start the sampling pump to extract soil water samples, filter them and introduce them into the flow-through comparison chamber, and use the reference electrode located in the comparison chamber to measure the reference value.

[0037] When calibration is required (e.g., due to redundancy discrepancies or reaching a fixed time interval), the system activates the sampling pump to extract soil water samples (1–10 mL), which are then filtered and introduced into the flow-through comparison chamber. The reference value E is measured using the reference electrode within the chamber. ref After measurement, the water sample was drained.

[0038] Step S4. Compare the actual measured value of the sensor with the expected response value calculated by the environmental coupling response model based on the environmental variables. When the deviation exceeds the second threshold, a second drift identification result is generated. Based on the environmental variables (T, W, S, I) collected in step S1, the expected response value E is calculated using the environmental coupling response model. true The actual measured value E of the sensor meas With E true The results are compared, and when the deviation exceeds the second threshold θ2, a second drift identification result is generated.

[0039] Step S5. Calculate the calibration coefficient based on the reference value and the first and second drift identification results, and perform temperature compensation and environmental compensation on the original measurement values ​​respectively.

[0040] Based on reference value E ref The measured value E of the sensor to be calibrated at the sampling time meas Calculate calibration coefficients Meanwhile, the original measured value E raw Subtract temperature compensation ΔE temp (T) Perform temperature compensation, and then calculate the environmental compensation amount ΔE based on the difference between the output of the environmental coupling model and the actual value. env .

[0041] Step S6. Calculate and output the final calibration value based on the calibration coefficient, temperature compensation amount, and environmental compensation amount.

[0042] According to the formula Calculate the final calibration value; where E raw The original measured value, ΔE temp ΔE is the temperature compensation amount. env Here, K represents the environmental compensation amount, and K is the calibration coefficient.

[0043] Step S7. Upload the final calibration value to the remote communication and management platform, and update the local operating parameters according to the platform instructions.

[0044] E out As a reliable monitoring data output, it also uploads raw data, calibration coefficients, health indices and other information to the remote platform; the platform decides whether to update local operating parameters based on preset strategies.

[0045] In a preferred embodiment, the calibration coefficient K is calculated according to the following formula: in, The reference base value, This represents the measurement value of the sensor to be calibrated at the sampling time.

[0046] In this embodiment, when performing micro water sample comparison, the reference electrode measures the reference value E. ref Simultaneously, the measured values ​​E of the sensor to be calibrated within the same time window are read. meas Divide the two to obtain the proportionality coefficient K; if K is not equal to 1, it indicates that the sensor has proportional drift; in subsequent measurements, the original sensor measurement value is multiplied by K for proportional correction, thereby eliminating the error caused by sensitivity changes; this coefficient is stored in the calibration parameter table and can be dynamically updated according to the next comparison result.

[0047] As a preferred embodiment, the final calibration value Calculate according to the following formula: in, The original measurement value of the sensor. This is the temperature compensation amount. For environmental compensation, This is the calibration coefficient.

[0048] In this embodiment, the temperature compensation function ΔE temp (T) can be a linear function (such as a·T+b) or a nonlinear function (such as a polynomial), and its parameters are obtained through calibration experiments of the sensor at different temperatures; during actual operation, ΔE is calculated based on the current temperature T. temp From the original measurement value E raw Subtracting it from the middle yields the temperature-corrected value; environmental compensation ΔE env Then, by comparing the expected value E output by the environment coupling model... true The measured value E after temperature correction temp_corr We obtain: ΔE env =E true -E temp_corr(Or the value can be obtained after filtering based on the model residuals); by adding this environmental compensation amount to the temperature-corrected value, the systematic deviation caused by the coupling of environmental factors such as salinity and water content can be eliminated.

[0049] In a preferred embodiment, the acquisition period of the reference value is dynamically adjusted based on the sensor's health index or historical drift trend; the health index is calculated as follows: in, For the first The deviation between the sensor measurement value and the reference value or the expected response value of the model in the comparison. For the rated range or specified reference value, This represents the number of comparisons.

[0050] In this embodiment, the system does not perform water sample comparisons at fixed time intervals, but rather adjusts them dynamically. For example, when the health index H is high (e.g., >0.95), it indicates that the sensor is in good condition, and the comparison cycle can be extended (e.g., from once a day to once a week). When H drops to a certain range (e.g., 0.8–0.95), the comparison cycle is increased (e.g., once a day). When H is below 0.7, not only is an alarm triggered, but real-time continuous monitoring may also be activated. The formula for calculating the health index is as shown above, where… The deviation between the sensor measurement and the reference value or the model expectation value in the i-th comparison is... The value is the rated range or reference value, and n is the number of comparisons. By dynamically adjusting the reference value acquisition cycle, the system power consumption (especially the energy consumption of the sampling pump) can be reduced while ensuring data accuracy.

[0051] In a preferred embodiment, the remote update performed by the remote communication and management platform includes at least one of the following: updating the parameters of the environmental coupling response model, modifying the first threshold or the second threshold, adjusting the data acquisition frequency of the sensor, or issuing updated values ​​of the calibration coefficients.

[0052] In this embodiment, the remote management platform sends update packages to the front-end nodes via OTA; specifically: Update environmental coupling response model parameters: The platform retrains the random forest or neural network model using long-term accumulated field monitoring data, compresses the new model coefficients or network weights and distributes them, and the front-end nodes receive them and replace the original model.

[0053] Correcting the first threshold for redundant comparison: The platform analyzes the historical difference statistical distribution of all nodes and dynamically adjusts θ1 to avoid frequent false alarms due to an overly strict threshold or missed alarms due to an overly lenient threshold.

[0054] Adjusting data collection frequency: Based on seasonal changes or user needs, the platform can issue instructions to change the collection cycle from 15 minutes to 30 minutes or 5 minutes.

[0055] Updated values ​​of calibration coefficients: In special circumstances, the platform can directly push a set of calibration coefficients to a specific sensor based on multi-node joint analysis, as initial values ​​or emergency corrections.

[0056] Through the aforementioned remote update strategy, the system possesses the ability to continuously evolve and adaptively optimize, maintaining long-term high-performance operation without on-site manual intervention.

[0057] Example 1 The system of this invention was deployed in a typical saline-alkali farmland in the North China Plain, with monitoring depths of 10cm (topsoil layer), 30cm (root distribution layer), and 50cm (subsoil layer); each monitoring node included a primary / secondary EC sensor, a primary / secondary pH sensor, and a Na+ sensor. + Ion-selective electrodes and temperature / moisture content sensors; a miniature soil water sampling and comparison unit is centrally installed in a protective well in the field, connected to soil water at various depths through pipelines, and automatically sampling and comparing at each depth in turn.

[0058] During system operation, raw data is collected every 15 minutes, and a micro water sample benchmark comparison is automatically performed at 2:00 AM daily (avoiding irrigation and rainfall interference); the remote management platform displays EC, pH, and Na at various depths in real time. + The platform displays the change curves of concentration and moisture content. When the EC at 10cm exceeds 8dS / m for three consecutive days and the health index H>0.9 (indicating sensor reliability), the platform automatically pushes guidance information to farmers' mobile phones stating "Topsoil salinity exceeds the standard, flood irrigation is recommended to suppress salt." At the same time, the platform retrains the environmental coupling model using data accumulated over three months and sends the updated model parameters to the front-end node via OTA, improving the temperature compensation accuracy by 15% in low-temperature winter environments. No data misjudgment due to sensor drift occurred throughout the spring and summer, and the number of manual on-site calibrations was reduced from once a month in the traditional scheme to only once a quarter (for replacing the reference electrode).

[0059] Example 2 The system of this invention was applied to a groundwater monitoring well network in a coastal area; the monitoring targets were groundwater conductivity and Cl. - Concentration, ion-selective electrode adjusted to Cl - Electrode selection and environmental coupling model incorporating tidal water level as an input variable; primary / auxiliary EC sensors and Cl are deployed in each monitoring well. - The electrode and the miniature soil water sampling and comparison unit were changed to a well pumping mode, and the reference electrode adopted a high-precision laboratory-grade EC / Cl. - Composite electrode.

[0060] The system collects data every 30 minutes and automatically performs a baseline comparison at 8:00 AM every day. When the groundwater conductivity is detected to rise sharply from 2 mS / cm to 5 mS / cm, and the difference between the main and auxiliary probes ΔE is less than 3% (indicating that the sensors are not faulty), the platform determines it to be a seawater intrusion event and immediately sends an early warning SMS to the water conservancy department and surrounding aquaculture farmers. At the same time, the system automatically deducts the influence of temperature (daily variation ±5℃) and tides (semi-diurnal cycle) on EC measurements through an environmental coupling model, avoiding false alarms caused by environmental fluctuations. After one year of continuous operation, only 5% of the sensors have a health index H below 0.7. The platform automatically generates a list of sensors that need maintenance and a well location distribution map for maintenance personnel to accurately replace them.

[0061] Example 3 Multiple monitoring points were set up around a salt mud stockpile of a chemical plant. Each point had monitoring nodes buried at depths of 20cm and 60cm, focusing on monitoring conductivity and pH value, and also adding Cl... - and SO4² - Ion-selective electrode; the miniature water sample collection unit is made of corrosion-resistant material, and the filter pore size is appropriately increased to accommodate samples with high suspended solids.

[0062] The system collects data every 10 minutes. When the EC value at a depth of 60cm at a certain monitoring point exceeds 15dS / m and the rise rate is greater than 0.5dS / m / h, and the main / auxiliary probe comparison is normal (ΔE<5%), the platform determines that an abnormal leakage has occurred and automatically sends a "start emergency pumping" command to the factory's central control system, linking the operation of surrounding pumping stations. Afterwards, the reliability of the data is confirmed by retrieving the historical health index of this node (H is always >0.85), avoiding the risk of erroneous equipment startup due to sensor drift. After six months of system operation, the platform automatically optimized the sampling frequency (from 10 minutes to 15 minutes, and 30 minutes at night) based on the accumulated leakage patterns, and reduced battery power consumption by issuing new strategies via OTA.

[0063] Example 4 In a coastal salinized wetland ecological restoration project area, three monitoring sections were set up along a direction perpendicular to the coastline, with five monitoring nodes set up in each section; the monitoring parameters included EC, pH, and Na. + Ca² + Soil moisture content was used to assess the effect of freshwater recharge on salt leaching; the micro soil water sampling and comparison unit was powered by solar energy and performed a baseline comparison every two days (salt content changes slowly in wetland environments).

[0064] The system monitored continuously for two years. In the first year, the raw data showed that EC decreased from 12 dS / m to 6 dS / m, but the temperature changed significantly during the same period. The environmental coupling model indicated that about 0.8 dS / m of the decrease was due to a false decrease caused by the increase in temperature, while the actual decrease in salinity was about 5.2 dS / m. In the second year, the data after model correction was more stable, and the assessment of the repair effect was more accurate. At the end of the project, the platform generated the health index change curve for each monitoring node, in which 90% of the sensors still had an H value higher than 0.75, proving that the self-calibration system of this invention can still work reliably for a long time in the highly corrosive and high-humidity coastal environment.

[0065] Example 5 In a saline-alkali land improvement demonstration area in Xinjiang Uygur Autonomous Region, the system of this invention is linked with underground drainage and drip irrigation salt leaching devices; the front-end monitoring nodes of the system are buried at depths of 20cm, 40cm and 60cm, and output the calibrated EC value and moisture content in real time; the remote management platform is connected to the irrigation controller and drainage pump controller via Modbus TCP protocol.

[0066] The platform is configured with the following rules: when EC at 20cm > 6dS / m and moisture content < 20%, drip irrigation for salt leaching is automatically initiated for 15 minutes; when EC at 60cm > 10dS / m and the rate of increase > 0.3dS / m / h, the underground drainage pump is automatically initiated for 30 minutes; all control commands are based on calibrated data; after three months of operation, a period of strong evaporation caused EC at 20cm to rise rapidly to 7.5dS / m, and the system automatically performed salt leaching, avoiding the harm of surface salt accumulation; subsequent manual sampling of soil salinity showed an error of less than 8% compared to the system output value, verifying the reliability of the self-calibrated data; the system achieves integrated closed-loop management of monitoring, diagnosis, decision-making, and control, significantly reducing the need for manual intervention.

[0067] Example 6 In a provincial-level saline-alkali land monitoring network, 50 sets of the system of this invention were deployed, covering 6 different climatic sub-regions. Each system operates independently and uploads calibrated data and health index to the provincial cloud platform via 4G. The platform aggregates the data from all nodes and establishes a unified data quality labeling system (Grade A: H≥0.9, data is reliable; Grade B: 0.7≤H<0.9, data is usable but requires attention; Grade C: H<0.7, data is unusable and requires maintenance).

[0068] Using this system, provincial researchers can quickly screen high-quality data for regional water-salt balance modeling. At the same time, the platform automatically generates a maintenance priority list for each node. After one year of operation, the platform analyzed the drift patterns of all nodes and found that the pH value of a certain model of main pH sensor decreased significantly faster than other models in alkaline soil. Therefore, it sent replacement recommendations to various cities and counties. This operation and maintenance strategy based on big data feedback has greatly improved the data quality and operational efficiency of the entire monitoring network.

[0069] Example 7 In a high-altitude, saline-alkali area in Qinghai, on-site maintenance is extremely difficult. The miniature soil water sampling and comparison unit of this invention is equipped with an automatic reference solution injection module. The system is set to perform reference electrode self-calibration every 30 days: automatically injecting standard EC solution (12.88 mS / cm, 25℃) and standard pH buffer solution (pH 9.18), recording the reference electrode readings, and automatically correcting internally if the accuracy requirements are not met. In this way, the reference electrode itself can maintain stability for one year without needing on-site replacement. Simultaneously, the main / auxiliary sensors utilize the E provided by the self-calibrated reference electrode monthly. ref The system was calibrated; it ran continuously for 18 months with a data validity rate of over 97%, and only required on-site intervention once due to a power failure, fully demonstrating the unattended operation capability of this invention in extremely remote areas.

[0070] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention; therefore, the embodiments should be regarded as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and thus all variations falling within the meaning and scope of the equivalents of the claims are intended to be included within the present invention; no reference numerals in the claims should be construed as limiting the scope of the claims.

[0071] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An intelligent self-calibration system for drift of a saline-alkali land monitoring sensor, characterized in that, include: A multi-parameter monitoring node is used to collect various soil parameters at the monitoring location, including at least electrical conductivity, pH value, ion concentration, temperature and moisture content. A redundancy comparison module, connected to the multi-parameter monitoring node, is used to cross-compare the output values ​​of the main sensor and the auxiliary sensor at the same monitoring location, and generate a first drift identification signal based on the comparison result. A miniature soil water sample collection and comparison unit includes a sampling pump, a filter, a flow-through comparison chamber, and a reference electrode disposed in the comparison chamber. This unit is used to extract water samples from the soil and obtain reference reference values ​​through the reference electrode in the flow-through comparison chamber. An environmental coupling drift identification module is used to acquire environmental variables, calculate the expected response value using an environmental coupling response model, and compare the actual measured value of the sensor with the expected response value to generate a second drift identification signal. The calibration compensation algorithm module is used to determine the calibration coefficient based on the first drift identification signal, the second drift identification signal and the reference value, and to perform temperature compensation, environmental compensation and proportional correction on the original sensor measurement value, and output the final calibration value. The remote communication and management platform communicates with the above modules and is used to receive the final calibration value and update the system operating parameters or processing logic remotely.

2. The system according to claim 1, characterized in that, The multi-parameter monitoring nodes include: a conductivity sensor forming a primary-secondary pairing, a pH sensor forming a primary-secondary pairing, at least one ion-selective electrode, and a temperature / moisture content sensor; and the paired sensors are different from each other in terms of range, structure, or measurement mechanism.

3. The system according to claim 1, characterized in that, The micro soil water sample collection and comparison unit also includes a standard solution storage and injection component, which is used to periodically inject standard solution into the flow-through comparison chamber to achieve online calibration of the reference electrode.

4. The system according to claim 1, characterized in that, The environmental coupled response model uses at least one of temperature, soil moisture content, salinity, and ionic strength as input parameters, and calculates the expected response value through a multinomial regression model, a random forest model, or a neural network model.

5. The system according to claim 1, characterized in that, The calibration compensation algorithm module is also used to calculate the health index of the sensor, wherein the health index H is determined according to the following formula: in, For the first The deviation between the sensor measurement value and the reference value or the expected response value of the model in the second comparison. Specify the sensor's rated range or a reference value. The number of comparisons included in the statistics; the calibration compensation algorithm module in the health index When the value falls below a set threshold, a maintenance alarm is issued through the remote communication and management platform.

6. A sensor drift self-calibration method for monitoring saline-alkali land, characterized in that, Includes the following steps: Collect real-time sensor measurements and environmental variables output from multi-parameter monitoring nodes; The measurements from the main sensor and the auxiliary sensor at the same monitoring location are compared. When the difference between the two exceeds a first threshold, a first drift identification result is generated. The sampling pump is started to extract soil water samples, which are then filtered and introduced into the flow-through comparison chamber. The reference value is measured using the reference electrode located in the comparison chamber. The actual measured value of the sensor is compared with the expected response value calculated by the environmental coupling response model based on the environmental variables. When the deviation exceeds the second threshold, a second drift identification result is generated. Based on the reference value and the first and second drift identification results, the calibration coefficient is calculated, and temperature compensation and environmental compensation are applied to the original measurement values ​​respectively. The final calibration value is calculated and output based on the calibration coefficient, temperature compensation, and environmental compensation. The final calibration value is uploaded to the remote communication and management platform, and the local operating parameters are updated according to the platform's instructions.

7. The method according to claim 6, characterized in that, The calibration coefficient K is calculated according to the following formula: in, The reference base value, This represents the measurement value of the sensor to be calibrated at the sampling time.

8. The method according to claim 6, characterized in that, The final calibration value Calculate according to the following formula: in, The original measurement value of the sensor. This is the temperature compensation amount. For environmental compensation, This is the calibration coefficient.

9. The method according to claim 6, characterized in that, The acquisition period of the reference value is dynamically adjusted based on the sensor's health index or historical drift trend; the health index is calculated as follows: in, For the first The deviation between the sensor measurement value and the reference value or the expected response value of the model in the second comparison. For the rated range or specified reference value, This represents the number of comparisons.

10. The method according to claim 6, characterized in that, The remote updates performed by the remote communication and management platform include at least one of the following: updating the parameters of the environmental coupling response model, modifying the first threshold or the second threshold, adjusting the data acquisition frequency of the sensor, or issuing updated values ​​of the calibration coefficients.

Images