An ec sensor automatic calibration method and system

Abstract

The application relates to an EC sensor automatic calibration method and system, which comprises the following steps: obtaining a calibration request, matching a calibration scheme; obtaining measurement data of the EC sensor in real time, triggering the EC sensor based on a calibration trigger condition; and when the calibration interval duration or the calibration trigger condition is met, automatically calibrating the EC sensor based on the calibration scheme. The application can intelligently decide to generate a calibration scheme through a calibration matching model according to a calibration request, has strong universality, can monitor a correction process in real time during the working period of the EC sensor, realizes intelligent and automatic calibration of the EC sensor, and effectively improves the correction efficiency and calibration precision of the EC sensor.

Description

Technical Field

[0001] This application relates to the field of EC sensors, and in particular to an automatic calibration method and system for EC sensors. Background Technology

[0002] An EC sensor (conductivity sensor) is a device used to measure the electrical conductivity of a solution. Electrical conductivity is a physical quantity that measures a substance's ability to conduct electricity. It has wide applications in many fields, including agriculture (for detecting the conductivity of soil and nutrient solutions), water quality monitoring (for detecting salinity in water bodies, etc.), and chemical experiments. Its measurement principle is based on the conductive properties of ions in a solution. When an electric field is applied across the solution, ions move directionally, generating an electric current. Conductivity is related to factors such as current intensity. The measurement accuracy of the sensor may decrease over time and with changes in the usage environment. Therefore, it is necessary to periodically calibrate the EC sensor to ensure that the conductivity value output by the sensor is accurate and reliable.

[0003] The existing automatic calibration process for EC sensors involves periodically immersing the sensor's electrode in a standard solution after a pre-set trigger condition is met. Then, the calibration coefficient is calculated by comparing the measured conductivity of the standard solution with a known standard value. Subsequent measurements are then corrected based on this calibration coefficient. However, calibrating EC sensors for different applications requires preparing specific standard solutions, making the preparation process cumbersome due to the need to consider various application scenarios. Furthermore, for precise calibration, the standard solution often needs to be replaced before each automatic calibration to avoid issues such as solution deterioration, inaccurate concentration, or contamination, resulting in high calibration costs.

[0004] Regarding the aforementioned technologies, the inventors believe that the existing EC sensor calibration preparation process is cumbersome and the calibration cost is too high. Summary of the Invention

[0005] To address the aforementioned issues, this application provides an automatic calibration method and system for EC sensors.

[0006] Firstly, this application provides an automatic calibration method for an EC sensor, employing the following technical solution:

[0007] An automatic calibration method for an EC sensor includes the following steps:

[0008] A calibration request is obtained, and a calibration scheme is generated accordingly. The calibration request includes basic information about the EC sensor, application environment information, and information about the target components to be measured. The calibration scheme includes a standard calibration solution formulation, a standard calibration environment, calibration trigger information, and calibration process information. The calibration trigger information includes the calibration interval duration and calibration trigger conditions.

[0009] Real-time acquisition of measurement data from the EC sensor; trigger verification of the EC sensor based on calibration trigger conditions.

[0010] When the calibration interval or calibration trigger condition is met, the EC sensor is automatically calibrated based on the calibration scheme.

[0011] Preferably, the real-time acquisition of measurement data from the EC sensor and the trigger verification of the EC sensor based on calibration trigger conditions specifically include the following steps:

[0012] The measurement data of the EC sensor is acquired in real time, and the measurement data is cleaned according to the calibration trigger conditions to obtain the measurement values ​​of the measurement items included in the calibration trigger conditions; the calibration trigger conditions include the calibration threshold of at least one measurement item of the EC sensor, and the calibration threshold includes the deviation calibration threshold, the rate of change calibration threshold, and the dynamic calibration coefficient K;

[0013] The difference between the measured value of the measurement item and the historical average measured value of the EC sensor within the measurement cycle is calculated to obtain the deviation difference value, and it is determined whether the deviation difference value exceeds the deviation calibration threshold.

[0014] If the time limit is not exceeded, the calibration trigger verification is considered to have passed.

[0015] If the value exceeds the limit, the measurement deviation of the measurement item is determined, and the rate of change of the measurement item is calculated according to the preset rate of change calculation formula. It is then determined whether the rate of change of the measurement item exceeds the rate of change calibration threshold. The specific rate of change calculation formula is: Rate of change = (current measurement value - previous measurement value) / previous measurement value × 100%.

[0016] If the time limit is not exceeded, the calibration trigger verification is deemed to have passed, and a sensor degradation warning is generated and sent to the management personnel.

[0017] If the value exceeds the limit, the calibration trigger verification is deemed to have failed, and the calibration trigger condition is met.

[0018] Preferably, the step of generating a sensor degradation alert and sending it to the administrator further includes: simultaneously generating a request to enable the dynamic deviation threshold when the administrator reads the sensor degradation alert; after the administrator agrees to enable it, calculating the dynamic deviation threshold of the EC sensor and replacing the deviation calibration threshold with the dynamic deviation threshold.

[0019] Preferably, the calculation of the dynamic deviation threshold of the EC sensor and the replacement of the deviation calibration threshold with the dynamic deviation threshold specifically includes the following steps:

[0020] Acquire measurement data from the EC sensor over the past unit measurement cycle and calculate the average difference between the measured values ​​over the past unit measurement cycle.

[0021] The standard deviation (SD) of the measured values ​​is calculated using a pre-set standard deviation formula, which is as follows:

[0022]

[0023] Where n is the total number of measurements in the EC sensor's measurement data within the past unit measurement cycle, and x i This refers to the measurement value of the i-th measurement within the measurement data of the EC sensor in the past unit measurement cycle;

[0024] The dynamic deviation threshold of the EC sensor is then calculated as K×SD, where K is the dynamic calibration coefficient in the calibration threshold, thereby determining the dynamic threshold range of the measured value. Replace the deviation calibration threshold with a dynamic deviation threshold.

[0025] Preferably, obtaining the calibration request and generating the calibration scheme specifically includes the following steps:

[0026] A calibration request is obtained, and multiple alternative schemes are generated by matching the pre-set calibration matching model. The calibration matching model is a convolutional neural network model, which is obtained by deep learning using historical working data and calibration data of EC sensors in different application fields as samples.

[0027] The score for each alternative solution is calculated using a pre-set comprehensive scoring formula; the comprehensive scoring formula is as follows:

[0028]

[0029] Among them, O j For the standard calibration solution cost of the j-th alternative, C j Let P be the average number of times the standard calibration solution for the j-th alternative is used, and O be the preset calibration cost baseline; j Let Q be the calibration duration for the j-th alternative, where P is the preset calibration duration reference; j Let H be the ambient temperature of the j-th alternative solution. j Let Z1 be the calibration ambient temperature for the j-th alternative, Q be the preset calibration temperature difference reference, Z1 be the cost coefficient, Z2 be the efficiency coefficient, and Z3 be the environmental coefficient. Z1, Z2, and Z3 are all set by the management personnel.

[0030] The alternative schemes are ranked and compared based on their scores, and the alternative scheme with the highest score is selected as the calibration scheme.

[0031] Preferably, the automatic calibration of the EC sensor based on the calibration scheme specifically includes the following steps:

[0032] Clean the electrodes of the EC sensor;

[0033] According to the calibration scheme, retrieve or configure the calibration solution formula, and start the EC sensor in a standard calibration environment to automatically calibrate and obtain the calibration coefficient based on the calibration process information.

[0034] The calibration scheme and calibration coefficients are input into a pre-set regression validation model for validation; the regression validation model is a linear regression model obtained through iterative training using historical calibration data.

[0035] If the verification passes, the automatic calibration is complete.

[0036] If the verification fails, an error message will be generated and sent to the administrator.

[0037] Preferably, it also includes: real-time acquisition of measurement data and calibration process data from the EC sensor, periodically packaged and sent to the calibration matching model for negative feedback iterative training.

[0038] Secondly, this application provides an automatic calibration system for EC sensors, which adopts the following technical solution:

[0039] An automatic calibration system for EC sensors includes:

[0040] The calibration scheme matching module is used to acquire calibration requests and generate calibration schemes. The calibration request includes basic information of the EC sensor, application environment information, and measurement target component information. The calibration scheme includes standard calibration solution formulation, standard calibration environment, calibration trigger information, and calibration process information. The calibration trigger information includes calibration interval duration and calibration trigger conditions.

[0041] The measurement and monitoring module is used to acquire measurement data from the EC sensor in real time and to perform trigger verification on the EC sensor based on calibration trigger conditions.

[0042] The automatic calibration module is used to automatically calibrate the EC sensor based on the calibration scheme when the calibration interval or calibration trigger condition is met.

[0043] Preferably, the automatic calibration module includes:

[0044] Electrode cleaning unit, used to clean the electrodes of EC sensor;

[0045] The automatic calibration unit is used to retrieve or configure the calibration solution formula according to the calibration scheme, and start the EC sensor to automatically calibrate in a standard calibration environment according to the calibration process information to obtain the calibration coefficient.

[0046] The calibration monitoring unit is used to input the calibration scheme and calibration coefficients into a preset regression verification model for verification. The regression verification model is a linear regression model obtained by iterative training using historical calibration data. If the verification passes, the calibration is automatically completed. If the verification fails, an error message is generated and sent to the management personnel.

[0047] Preferably, the scheme matching module includes:

[0048] The scheme matching unit is used to obtain calibration requests and generate multiple alternative schemes by matching them through a pre-set calibration matching model. The calibration matching model is a convolutional neural network model, which is obtained by deep learning using historical working data and calibration data of EC sensors in different application fields as samples.

[0049] The scheme scoring unit is used to calculate the scheme score of each alternative scheme using a preset comprehensive scoring formula; the comprehensive scoring formula is as follows:

[0050]

[0051] Among them, O j For the standard calibration solution cost of the j-th alternative, C j Let P be the average number of times the standard calibration solution for the j-th alternative is used, and O be the preset calibration cost baseline; j Let Q be the calibration duration for the j-th alternative, where P is the preset calibration duration reference; j Let H be the ambient temperature of the j-th alternative solution. j Let Z1 be the calibration ambient temperature for the j-th alternative, Q be the preset calibration temperature difference reference, Z1 be the cost coefficient, Z2 be the efficiency coefficient, and Z3 be the environmental coefficient. Z1, Z2, and Z3 are all set by the management personnel.

[0052] The scheme selection unit is used to rank and compare the various candidate schemes based on the scheme score, and select the candidate scheme with the highest scheme score as the calibration scheme.

[0053] In summary, this application includes at least one of the following beneficial technical effects:

[0054] 1. Based on the calibration request, a calibration matching model intelligently generates a calibration scheme, which is highly universal and adaptable to various application scenarios. It eliminates the need for extensive manpower and resources for calibration preparation, achieving accurate, efficient, and automatic calibration of sensors. During the actual use of EC sensors, it can determine whether the sensor is faulty based on the current measurement data and environmental conditions. For EC sensors with abnormal data, it directly performs calibration, ensuring stable measurement operations. When EC sensors need calibration, it automatically corrects the sensors according to the calibration scheme, reducing manual intervention, improving calibration consistency, and effectively improving the calibration efficiency and accuracy of EC sensors.

[0055] 2. By using a calibration matching model to generate multiple alternative solutions that meet the calibration request, the alternative solutions are then scored based on factors such as calibration cost, solution preservation difficulty, calibration efficiency, and the difference between the calibration environment and the application environment. The optimal calibration solution is then determined, thereby achieving more accurate calibration and improving measurement accuracy.

[0056] 3. By cleaning the data according to the calibration trigger conditions, we can focus on the measurement items that are truly related to calibration, which helps to remove the interference of irrelevant data. Then, we perform dual verification of deviation and rate of change. By first judging the difference between the measured value and the historical average measured value, and then checking the rate of change, we have built a dual verification mechanism. This mechanism can more comprehensively detect abnormalities in the measurement data, accurately locate the problem type, respond to anomalies in a timely manner, and calibrate EC sensors with anomalies in advance, so as to ensure the stable operation of EC sensor measurements. Attached Figure Description

[0057] Figure 1 This is a flowchart of an automatic calibration method for an EC sensor according to an embodiment of this application;

[0058] Figure 2 This is a flowchart of the method for generating a calibration scheme in the embodiments of this application;

[0059] Figure 3 This is a flowchart of the method for triggering verification of the EC sensor in this application embodiment;

[0060] Figure 4 This is a flowchart of the method for calculating the dynamic deviation threshold of the EC sensor in an embodiment of this application;

[0061] Figure 5 This is a flowchart of the method for automatically calibrating an EC sensor in an embodiment of this application;

[0062] Figure 6 This is a system block diagram of an automatic calibration system for an EC sensor according to an embodiment of this application.

[0063] Explanation of reference numerals in the attached diagram: 1. Scheme matching module; 11. Scheme matching unit; 12. Scheme scoring unit; 13. Scheme selection unit; 2. Measurement and monitoring module; 3. Automatic calibration module; 31. Electrode cleaning unit; 32. Automatic calibration unit; 33. Calibration monitoring unit. Detailed Implementation

[0064] The following is in conjunction with the appendix Figure 1-6 This application will be described in further detail.

[0065] This application discloses an automatic calibration method for an EC sensor. (Refer to...) Figure 1 An automatic calibration method for an EC sensor includes the following steps:

[0066] S1. Obtain a calibration request and generate a calibration scheme; the calibration request includes basic EC sensor information, application environment information, and target component measurement information; the calibration scheme includes standard calibration solution formulation, standard calibration environment, calibration trigger information, and calibration process information; the calibration trigger information includes calibration interval duration and calibration trigger conditions.

[0067] S2. Acquire measurement data from the EC sensor in real time and perform trigger verification on the EC sensor based on calibration trigger conditions;

[0068] S3. When the calibration interval or calibration trigger condition is met, the EC sensor is automatically calibrated based on the calibration scheme. Through the above steps, a calibration scheme is generated intelligently through the calibration matching model based on the calibration request. This method is highly universal and adaptable to various application scenarios, eliminating the need for extensive manpower and resources for calibration preparation. It achieves accurate, efficient, and automatic calibration of the sensor. During the actual use of the EC sensor, the system can determine whether the sensor is faulty based on the current measurement data and environmental conditions. EC sensors with abnormal data are directly calibrated, ensuring stable measurement operations. When EC sensor calibration is required, it is automatically corrected according to the calibration scheme, reducing manual intervention, improving calibration consistency, and effectively improving the calibration efficiency and accuracy of the EC sensor.

[0069] Reference Figure 2 The process of obtaining a calibration request and generating a calibration scheme specifically includes the following steps:

[0070] A1. Generate alternative solutions: Obtain a calibration request and generate multiple alternative solutions by matching them with a pre-set calibration matching model. The calibration matching model is a convolutional neural network model, which is obtained by deep learning using historical working data and calibration data of EC sensors in different application fields as samples.

[0071] A2. Calculate the score of each alternative scheme: Calculate the score of each alternative scheme using a pre-set comprehensive scoring formula; the comprehensive scoring formula is as follows:

[0072]

[0073] Among them, O j For the standard calibration solution cost of the j-th alternative, C j Let P be the average number of times the standard calibration solution for the j-th alternative is used, and O be the preset calibration cost baseline; j Let Q be the calibration duration for the j-th alternative, where P is the preset calibration duration reference; j Let H be the ambient temperature of the j-th alternative solution. j Let Z1 be the calibration ambient temperature for the j-th alternative, Q be the preset calibration temperature difference reference, Z1 be the cost coefficient, Z2 be the efficiency coefficient, and Z3 be the environmental coefficient. Z1, Z2, and Z3 are all set by the management personnel.

[0074] A3. Calibration Scheme Selection: Each candidate scheme is ranked and compared based on its score, and the scheme with the highest score is selected as the calibration scheme. A calibration matching model generates multiple candidate schemes that meet the calibration request. These schemes are then evaluated based on factors such as calibration cost (solution cost), solution preservation difficulty (easy solution preservation naturally reduces per-schedule cost), calibration efficiency, and the difference between the calibration environment and the application environment (calibration convenience). The optimal calibration scheme is then determined, resulting in more accurate calibration and improved measurement accuracy.

[0075] In addition, during automatic calibration of the EC sensor, measurement data and calibration process data are collected in real time and periodically packaged and sent to the calibration matching model for negative feedback iterative training. This helps to further improve and optimize the calibration matching model, simplifying the calibration preparation process while effectively improving the accuracy and sophistication of the calibration scheme matching.

[0076] Reference Figure 3 The above-mentioned real-time acquisition of measurement data from the EC sensor and the trigger verification of the EC sensor based on calibration trigger conditions specifically include the following steps:

[0077] B1. Data Acquisition and Cleaning: Real-time acquisition of measurement data from the EC sensor, and data cleaning of the measurement data according to the calibration trigger conditions to obtain the measurement values ​​of the measurement items included in the calibration trigger conditions; the calibration trigger conditions include calibration thresholds for at least one measurement item of the EC sensor, and the calibration thresholds include deviation calibration thresholds, rate of change calibration thresholds, and dynamic calibration coefficient K;

[0078] B2. Calculate the difference between the measured value of the measurement item and the historical average measured value of the EC sensor within the measurement cycle to obtain the deviation difference, and determine whether the deviation difference exceeds the deviation calibration threshold.

[0079] B3. If the test result is within the specified time, the calibration trigger verification is considered successful.

[0080] B4. If it exceeds the limit, the measurement deviation of the measurement item is determined, and the change rate of the measurement item is calculated according to the preset change rate calculation formula to determine whether the change rate of the measurement item exceeds the change rate calibration threshold. The change rate calculation formula is as follows: Change rate = (current measurement value - previous measurement value) / previous measurement value × 100%.

[0081] B5. Calibration trigger verification passed: If the time limit is not exceeded, the calibration trigger verification is deemed to have passed, and a sensor degradation prompt is generated and sent to the management personnel.

[0082] B6. Calibration trigger condition met: If exceeded, the calibration trigger verification is deemed failed; otherwise, the calibration trigger condition is met. Data cleaning based on the calibration trigger condition allows focus on truly calibration-related measurements, helping to remove irrelevant data interference. A dual verification mechanism is then implemented, checking both deviation and rate of change. This mechanism first determines the difference between the measured value and the historical average, then examines the rate of change. This comprehensive approach detects anomalies in the measurement data, accurately identifies problem types, responds promptly, and allows for early calibration of EC sensors with anomalies, ensuring stable EC sensor measurements.

[0083] Additionally, the generation of sensor degradation alerts sent to management personnel includes: simultaneously generating a request to enable dynamic deviation thresholds when management reads the alert. After management approval, the dynamic deviation threshold for the EC sensor is calculated and replaced with the deviation calibration threshold. In cases where calibration trigger verification passes but measurement deviations are detected, a sensor degradation alert is generated and sent to management. This allows management to promptly understand the sensor's operating status and apply dynamic deviation thresholds to degraded sensors. To save costs and continue use, monitoring capabilities are adaptively improved, dynamically adapting to sensor changes, optimizing trigger accuracy, reducing unnecessary calibration operations, and extending sensor lifespan. Furthermore, management can make flexible decisions based on actual conditions. This flexibility allows for the selection of the most suitable calibration triggering strategy when facing different sensors, different measurement tasks, and different resource constraints. For example, in some chemical experiment monitoring scenarios with extremely high accuracy requirements, dynamic deviation thresholds may be enabled when the sensor shows slight signs of degradation to ensure the accuracy of experimental data; while in some environmental monitoring scenarios with slightly lower accuracy requirements, they can be enabled only after the sensor has degraded to a certain extent, balancing accuracy and cost.

[0084] Reference Figure 4 The calculation of the dynamic deviation threshold of the EC sensor and the replacement of the deviation calibration threshold with the dynamic deviation threshold specifically includes the following steps:

[0085] C1. Data Acquisition: Obtain measurement data from the EC sensor over the past unit measurement cycle, and calculate the average difference between the measured values ​​over the past unit measurement cycle.

[0086] C2. Calculate the standard deviation of the measured values: Calculate the standard deviation SD of the measured values ​​using a preset standard deviation formula, wherein the standard deviation formula is as follows:

[0087]

[0088] Where n is the total number of measurements in the EC sensor's measurement data within the past unit measurement cycle, and x i This refers to the measurement value of the i-th measurement within the measurement data of the EC sensor in the past unit measurement cycle;

[0089] C3. Calculate the dynamic deviation threshold: Then calculate the dynamic deviation threshold of the EC sensor as K×SD, where K is the dynamic calibration coefficient in the calibration threshold, and thus determine the dynamic threshold range of the measured value. The dynamic deviation threshold is used to replace the deviation calibration threshold. Upon detection of sensor degradation and a decision by management, the dynamic deviation threshold is replaced with the deviation calibration threshold through the steps described above to ensure the accuracy of experimental data and achieve intelligent automatic calibration.

[0090] Reference Figure 5 The above-mentioned automatic calibration of EC sensors based on the calibration scheme specifically includes the following steps:

[0091] D1. Electrode Cleaning: Clean the electrodes of the EC sensor.

[0092] D2. Automatic calibration: Retrieve or configure the calibration solution formula according to the calibration scheme, and start the EC sensor in the standard calibration environment to obtain the calibration coefficient according to the calibration process information.

[0093] D3. Calibration Monitoring: Input the calibration scheme and calibration coefficients into a pre-set regression verification model for verification; the regression verification model is a linear regression model obtained through iterative training using historical calibration data.

[0094] D4. If the verification passes, the automatic calibration is complete.

[0095] D5. Error Report Generation: If the verification fails, an error message is generated and sent to the management personnel. After completing sensor calibration, a regression verification model predicts an expected calibration coefficient range based on the calibration scheme. If the actual calibration coefficients exceed this range, a problem may exist, and an error message is generated and sent to the management personnel. Through the above steps, the sensor calibration work can be monitored and evaluated in real time, potential calibration errors can be detected in advance, and troubleshooting guidance can be provided for EC sensors with calibration anomalies. This ensures stable EC sensor detection, helps improve the measurement quality of EC sensors, prevents erroneous data from being used in subsequent measurements, and effectively improves the calibration efficiency and accuracy of EC sensors.

[0096] To illustrate with a detailed example, after cleaning the electrodes, the electrodes of the EC sensor to be calibrated are placed in a standard calibration solution, and the EC sensor is activated for measurement. In this example, the measurement item is the conductivity of the solution, so the EC sensor begins measuring the conductivity of the standard solution. This measurement process typically lasts for a period of time to obtain stable and reliable measurement data. During the measurement process, the sensor acquires data according to its own measurement principles (such as measuring the potential difference between electrodes). Let the measured conductivity of the standard solution be EC1, and the known conductivity of the standard solution be EC2, then the calibration coefficient F = EC2 / EC1. This coefficient will be used to correct subsequent actual measurement data. In the subsequent actual measurement, let the uncorrected measurement data be EC3, then the corrected measurement data EC4 = EC3 * F. This is the existing automatic calibration process for EC sensors. The above steps, after calibration, involve inputting the calibration scheme and calibration coefficients into a pre-set regression validation model. This regression validation model is a linear regression model within machine learning. A linear regression model can predict the calibration coefficients under normal conditions by fitting a linear relationship between the calibration coefficients and factors such as the concentration of the standard solution and temperature. After each calibration of the EC sensor, the new calibration coefficients, along with the corresponding standard solution information and environmental parameters (calibration coefficients and calibration scheme), are input into the trained machine learning model. The model calculates and judges based on the input data. The model predicts an expected range of calibration coefficients. If the actual calibration coefficients exceed this range, there may be a problem, meaning the validation has failed.

[0097] This application also discloses an automatic calibration system for EC sensors. (See also...) Figure 6 An automatic calibration system for EC sensors, comprising:

[0098] The scheme matching module 1 is used to acquire calibration requests and generate calibration schemes by matching them. The calibration request includes basic information of the EC sensor, application environment information, and measurement target component information. The calibration scheme includes standard calibration solution formulation, standard calibration environment, calibration trigger information, and calibration process information. The calibration trigger information includes calibration interval duration and calibration trigger conditions.

[0099] Measurement and monitoring module 2 is used to acquire measurement data from the EC sensor in real time and to perform trigger verification on the EC sensor based on calibration trigger conditions;

[0100] Automatic calibration module 3 is used to automatically calibrate the EC sensor based on the calibration scheme when the calibration interval or calibration trigger condition is met.

[0101] Additionally, a negative feedback module 4 is included. This module collects EC sensor measurement data and calibration process data in real time during automatic EC sensor calibration, and periodically packages and sends this data to the calibration matching model for iterative negative feedback training. Based on calibration requests, the calibration matching model intelligently generates calibration schemes, offering strong versatility and adaptability to various application scenarios. It eliminates the need for extensive manpower and resources for calibration preparation, achieving precise and efficient automatic sensor calibration. Furthermore, during actual use of the EC sensor, it can determine if the sensor is faulty based on current measurement data and environmental conditions. EC sensors with abnormal data are directly calibrated, ensuring stable measurement operations. When EC sensor calibration is required, it automatically corrects the sensor according to the calibration scheme, reducing manual intervention, improving calibration consistency, and effectively enhancing the calibration efficiency and accuracy of the EC sensor.

[0102] Reference Figure 6 The scheme matching module 1 includes:

[0103] The scheme matching unit 11 is used to obtain calibration requests and generate multiple alternative schemes by matching through a pre-set calibration matching model. The calibration matching model is a convolutional neural network model, which is obtained by deep learning using historical working data and calibration data of EC sensors in different application fields as samples.

[0104] The scheme scoring unit 12 is used to calculate the scheme score of each candidate scheme using a preset comprehensive scoring calculation formula; the comprehensive scoring calculation formula is as follows:

[0105]

[0106] Among them, O j For the standard calibration solution cost of the j-th alternative, C jLet P be the average number of times the standard calibration solution for the j-th alternative is used, and O be the preset calibration cost baseline; j Let Q be the calibration duration for the j-th alternative, where P is the preset calibration duration reference; j Let H be the ambient temperature of the j-th alternative solution. j Let Z1 be the calibration ambient temperature for the j-th alternative, Q be the preset calibration temperature difference reference, Z1 be the cost coefficient, Z2 be the efficiency coefficient, and Z3 be the environmental coefficient. Z1, Z2, and Z3 are all set by the management personnel.

[0107] The scheme selection unit 13 is used to rank and compare the various candidate schemes based on the scheme score, and select the candidate scheme with the highest scheme score as the calibration scheme. Multiple candidate schemes that meet the calibration request are generated based on the calibration matching model. Then, each candidate scheme is scored based on factors such as calibration cost (solution cost), solution preservation difficulty (easy solution preservation naturally reduces the cost per attempt), calibration efficiency, and the difference between the calibration environment and the application environment (calibration convenience), thereby determining the optimal calibration scheme, achieving more accurate calibration, and improving measurement accuracy.

[0108] Reference Figure 6 The automatic calibration module 3 includes:

[0109] Electrode cleaning unit 31 is used to clean the electrodes of the EC sensor.

[0110] Automatic calibration unit 32 is used to retrieve or configure calibration solution formula according to calibration scheme, and start EC sensor to perform automatic calibration in standard calibration environment according to calibration process information to obtain calibration coefficient;

[0111] The calibration monitoring unit 33 is used to input the calibration scheme and calibration coefficients into a pre-set regression verification model for verification. The regression verification model is a linear regression model obtained through iterative training using historical calibration data. If the verification passes, the calibration is automatically completed; if the verification fails, an error message is generated and sent to the management personnel. By cleaning the data according to the calibration trigger conditions, the focus can be placed on the measurement items that are truly related to calibration, which helps to remove interference from irrelevant data. Then, a dual verification mechanism is constructed by first judging the difference between the measured value and the historical average measured value, and then checking the rate of change. This mechanism can more comprehensively detect abnormalities in the measurement data, accurately locate the problem type, respond to anomalies in a timely manner, and calibrate EC sensors with anomalies in advance, ensuring the stable operation of EC sensor measurements.

[0112] This application also discloses a computer-readable storage medium that stores a computer program that can be loaded by a processor and executed as described above. The computer-readable storage medium includes, for example, various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0113] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on these embodiments, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can still combine, add, delete, or otherwise adjust the features of the various embodiments of the present invention according to the circumstances without conflict or creative effort, thereby obtaining different technical solutions that do not fundamentally depart from the concept of the present invention. These technical solutions also fall within the scope of protection of the present invention.

Claims

1. An EC sensor automatic calibration method, characterized by, The method comprises the following steps: obtaining a calibration request and matching a generated calibration scheme; the calibration request comprises EC sensor basic information, application environment information and measurement target component information, and the calibration scheme comprises standard calibration solution formula, standard calibration environment, calibration trigger information and calibration process information; the calibration trigger information comprises calibration interval duration and calibration trigger condition; real-time acquisition of measurement data of the EC sensor, and trigger verification of the EC sensor based on the calibration trigger condition; when the calibration interval duration or the calibration trigger condition is met, automatic calibration of the EC sensor based on the calibration scheme; the real-time acquisition of the measurement data of the EC sensor, and the trigger verification of the EC sensor based on the calibration trigger condition specifically comprises the following steps: real-time acquisition of the measurement data of the EC sensor, and data cleaning of the measurement data according to the calibration trigger condition to obtain the measurement value of the measurement item included in the calibration trigger condition; the calibration trigger condition comprises a calibration threshold of at least one measurement item of the EC sensor, and the calibration threshold comprises a deviation calibration threshold, a change rate calibration threshold and a dynamic calibration coefficient K; calculating the difference between the measurement value of the measurement item and the historical average measurement value of the EC sensor in the measurement period to obtain a deviation difference, and judging whether the deviation difference exceeds the deviation calibration threshold; if not, it is determined that the calibration trigger verification is passed; if it exceeds, it is determined that the measurement deviation of the measurement item, and then a change rate of the measurement item is calculated according to a pre-set change rate calculation formula; the change rate calculation formula is specifically: change rate = (this time measurement value - last time measurement value) / last time measurement value ×100%； if it does not exceed, it is determined that the calibration trigger verification is passed, and a sensor degradation prompt is generated and sent to the administrator; if it exceeds, it is determined that the calibration trigger verification is not passed, and the calibration trigger condition is met; the generation of the sensor degradation prompt sent to the administrator further comprises: when the administrator reads the sensor degradation prompt, a request for enabling a dynamic deviation threshold is generated synchronously, and after the administrator agrees to start, a dynamic deviation threshold of the EC sensor is calculated, and the dynamic deviation threshold is used to replace the deviation calibration threshold; the calculation of the dynamic deviation threshold of the EC sensor and the replacement of the dynamic deviation threshold with the deviation calibration threshold specifically comprises the following steps: acquiring measurement data of the EC sensor for a past unit measurement period, calculating an average difference of the measurement values for the past unit measurement period ; the standard deviation SD of the measurement value is calculated by a pre-set standard deviation formula; the standard deviation formula is specifically: ； wherein n is the total number of measurements in the measurement data of the EC sensor in the past unit measurement period, is the measurement value of the i-th measurement in the measurement data of the EC sensor in the past unit measurement period. Further, the dynamic deviation threshold of the EC sensor is calculated as KxSD, wherein K is a dynamic calibration coefficient in the calibration threshold, and further the dynamic threshold range of the measurement value is determined as (X-KxSD, X+KxSD), and the dynamic deviation threshold is used to replace the deviation calibration threshold. -KxSD, +KxSD).

2. The method of claim 1, wherein: the acquisition of the calibration request and the matching generation of the calibration scheme specifically comprises the following steps: obtaining a calibration request, matching a plurality of alternative schemes by a pre-set calibration matching model, the calibration matching model is a convolutional neural network model, and the calibration matching model is obtained by deep learning of historical working data and calibration data of EC sensors in different application fields as samples; a scheme score of each alternative scheme is calculated by a pre-set comprehensive score calculation formula; the comprehensive score calculation formula is specifically: Y= ； wherein, is the standard calibration solution cost of the jth alternative, is the average use frequency of the calibration solution of the jth alternative, and O is a pre-set calibration cost benchmark; is the calibration duration of the jth alternative, and P is a pre-set calibration duration benchmark; is the application environment temperature of the jth alternative, is the calibration environment temperature of the jth alternative, and Q is a pre-set calibration temperature difference benchmark; is a cost coefficient, is an efficiency coefficient, is an environment coefficient, and , , are all set by the administrator; each alternative scheme is sorted and compared based on the scheme score, and the alternative scheme with the highest scheme score is selected as the calibration scheme.

3. The method of claim 1, wherein, the automatic calibration of the EC sensor based on the calibration scheme specifically comprises the following steps: cleaning treatment of the electrode of the EC sensor; According to the calibration scheme, the calibration solution formula is called or configured, the EC sensor is started for automatic calibration in the standard calibration environment according to the calibration process information to obtain a calibration coefficient; The calibration scheme and the calibration coefficient are input into a preset regression verification model for verification; the regression verification model is a linear regression model obtained by iterative training of historical calibration data; If the verification is passed, the automatic calibration is completed; If the verification is not passed, error information is generated and sent to the management personnel.

4. The method of claim 2, wherein, Also includes: Real-time acquisition of the measurement data and the calibration process data of the EC sensor, and periodic packaging and sending to the calibration matching model for negative feedback iterative training.

5. An EC sensor auto-calibration system, characterized by, Includes: The scheme matching module (1) is used for obtaining a calibration request and matching to generate a calibration scheme; the calibration request includes EC sensor basic information, application environment information and measurement target component information, the calibration scheme includes a standard calibration solution formula, a standard calibration environment, calibration trigger information and calibration process information; the calibration trigger information includes a calibration interval duration and a calibration trigger condition; The measurement monitoring module (2) is used for real-time acquisition of the measurement data of the EC sensor and trigger verification of the EC sensor based on the calibration trigger condition; The automatic calibration module (3) is used for automatic calibration of the EC sensor based on the calibration scheme when the calibration interval duration or the calibration trigger condition is met; The measurement monitoring module (2) real-time acquisition of the measurement data of the EC sensor and trigger verification of the EC sensor based on the calibration trigger condition specifically includes the following steps: Real-time acquisition of the measurement data of the EC sensor, and data cleaning of the measurement data according to the calibration trigger condition to obtain the measurement value of the measurement item included in the calibration trigger condition; the calibration trigger condition includes a calibration threshold of at least one measurement item of the EC sensor, and the calibration threshold includes a deviation calibration threshold, a change rate calibration threshold and a dynamic calibration coefficient K; The difference between the measurement value of the measurement item and the historical average measurement value of the EC sensor in the measurement period is calculated to obtain a deviation difference, and it is judged whether the deviation difference exceeds the deviation calibration threshold; If not, it is determined that the calibration trigger verification is passed; If it exceeds, it is determined that the measurement deviation of the measurement item, and the change rate of the measurement item is calculated according to a preset change rate calculation formula; the change rate calculation formula is specifically: change rate = (this time measurement value - last time measurement value) / last time measurement value ×100%； If it does not exceed, it is determined that the calibration trigger verification is passed, and a sensor degradation prompt is generated and sent to the management personnel; If it exceeds, it is determined that the calibration trigger verification is not passed, and the calibration trigger condition is met; The generation of the sensor degradation prompt sent to the management personnel also includes: when the management personnel reads the sensor degradation prompt, a request for enabling a dynamic deviation threshold is generated synchronously, after the management personnel agrees to start, the dynamic deviation threshold of the EC sensor is calculated, and the dynamic deviation threshold is used to replace the deviation calibration threshold The calculation of the dynamic deviation threshold of the EC sensor and the replacement of the dynamic deviation threshold with the deviation calibration threshold specifically includes the following steps: acquiring measurement data of the EC sensor for a past unit measurement period, calculating an average difference of the measurement values for the past unit measurement period ; The standard deviation SD of the measurement value is calculated by a preset standard deviation formula; the standard deviation formula is specifically: ； wherein n is the total number of measurements in the measurement data of the EC sensor in the past unit measurement period, is the measurement value of the i-th measurement in the measurement data of the EC sensor in the past unit measurement period. Further, the dynamic deviation threshold of the EC sensor is calculated as KxSD, wherein K is a dynamic calibration coefficient in the calibration threshold, and further the dynamic threshold range of the measurement value is determined as (X-KxSD, X+KxSD), and the dynamic deviation threshold replaces the deviation calibration threshold 4. -KxSD, +KxSD).

6. An EC sensor automatic calibration system according to claim 5, wherein, The automatic calibration module (3) comprises: An electrode cleaning unit (31) configured to clean electrodes of the EC sensor; An automatic calibration unit (32) configured to retrieve or configure a calibration solution formula according to a calibration scheme, and to start the EC sensor to perform automatic calibration to obtain a calibration coefficient according to calibration process information in a standard calibration environment; A calibration monitoring unit (33) configured to input the calibration scheme and the calibration coefficient into a pre-set regression verification model for verification, the regression verification model being a linear regression model obtained by iterative training of historical calibration data, and if the verification is passed, the automatic calibration is completed, and if the verification is not passed, error information is generated and sent to a manager.

7. The automatic calibration system for an EC sensor of claim 5, wherein, The scheme matching module (1) comprises: A scheme matching unit (11) configured to obtain a calibration request, and to match a plurality of candidate schemes through a pre-set calibration matching model, the calibration matching model being a convolutional neural network model obtained through deep learning of historical working data and calibration data of EC sensors in different application fields as samples; A scheme scoring unit (12) configured to calculate a scheme score of each candidate scheme through a pre-set comprehensive scoring calculation formula, the comprehensive scoring calculation formula being specifically: Y= ； wherein, is the standard calibration solution cost of the jth alternative solution, is the average use frequency of the standard calibration solution of the jth alternative solution, and O is a pre-set calibration cost benchmark; is the calibration duration of the jth alternative solution, and P is a pre-set calibration duration benchmark; is the application environment temperature of the jth alternative solution, is the calibration environment temperature of the jth alternative solution, and Q is a pre-set calibration temperature difference benchmark; is a cost coefficient, is an efficiency coefficient, is an environment coefficient, and , , are all set by the administrator; the solution selection unit (13) is configured to sort and compare each alternative solution based on the solution score, and select the alternative solution with the highest solution score as the calibration solution.

Images