EC sensor automatic calibration method and system

Abstract

The invention relates to an EC sensor automatic calibration method and system, and the method comprises the following steps: obtaining a calibration request, and carrying out the matching to generate a calibration scheme; acquiring measurement data of the EC sensor in real time, and performing trigger verification on the EC sensor based on the calibration trigger condition; and automatically calibrating the EC sensor based on the calibration scheme when the calibration interval duration is met or the calibration trigger condition is met. According to the method, the calibration scheme can be intelligently decided and generated through the calibration matching model according to the calibration request, the universality is high, the EC sensor is monitored in real time in the calibration process in the working period, intelligent automatic calibration of the EC sensor is achieved, and the effect of effectively improving the calibration efficiency and calibration precision of the EC sensor is achieved.

Description

Technical Field

[0001] The present application relates to the field of EC sensors, and in particular to an EC sensor automatic calibration method and system. Background Art

[0002] EC sensor (conductivity sensor) is a device used to measure the conductivity of a solution. Conductivity is a physical quantity that measures the ability of a substance to conduct electricity. It is widely used in many fields, such as agriculture (used to detect the conductivity of soil and nutrient solution), water quality monitoring (detecting the salt content of water, etc.), and chemical experiments. Its measurement principle is based on the conductive properties of ions in the solution. When an electric field is applied to both ends of the solution, the ions will move in a directional manner, thereby generating current. The conductivity is related to factors such as the current intensity. With the change of time and the use environment, the measurement accuracy of the sensor may decrease, so it is necessary to calibrate the EC sensor regularly to ensure that the conductivity value output by the sensor is true and reliable.

[0003] The existing automatic calibration process of EC sensors is as follows: regularly or after the preset trigger conditions are met, the electrodes of the EC sensor are immersed in the standard solution, and then the calibration coefficient is calculated by comparing the measured conductivity value of the standard solution with the known standard value. In subsequent actual measurements, the measurement data is corrected according to this calibration coefficient. When calibrating existing EC sensors for different application scenarios, it is necessary to configure exclusive standard solutions for them. The preliminary preparations need to consider various application scenarios, resulting in cumbersome preparations. At the same time, when the EC sensor needs to be accurately calibrated, in order to avoid deterioration of the solution, inaccurate concentration or contamination, it is often necessary to replace the standard solution before each automatic calibration, resulting in high calibration costs.

[0004] With respect to the above-mentioned related 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] In order to solve the above problems, the present application provides an EC sensor automatic calibration method and system.

[0006] In a first aspect, the present application provides an EC sensor automatic calibration method, which adopts the following technical solution:

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

[0008] Obtain a calibration request and match and generate a calibration scheme; the calibration request includes basic information of the EC sensor, 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;

[0009] Acquire the measurement data of the EC sensor in real time and perform trigger verification on the EC sensor based on the calibration trigger conditions;

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

[0011] Preferably, 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 include the following steps:

[0012] Acquire measurement data of the EC sensor in real time, and perform data cleaning on the measurement data according to a calibration trigger condition to obtain a measurement value of a 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;

[0013] Calculate the difference between the measured value of the measurement item and the historical average measured value of the EC sensor in the measurement cycle to obtain a deviation difference, and determine whether the deviation difference exceeds the deviation calibration threshold;

[0014] If it does not exceed, the calibration trigger verification is determined to have passed;

[0015] If it exceeds, the measurement deviation of the measurement item is determined, and then 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 specifically: change rate = (this measurement value - last measurement value) / last measurement value × 100%;

[0016] If it does not exceed, the calibration trigger verification is determined to have passed, and a sensor degradation prompt is generated and sent to the management staff;

[0017] If it exceeds, it is determined that the calibration trigger verification has not passed and the calibration trigger condition is met.

[0018] Preferably, the generating of the sensor degradation prompt and sending it to the management personnel also includes: when the management personnel reads the sensor degradation prompt, synchronously generating a request to enable the dynamic deviation threshold, after the management personnel agrees to start, calculating the dynamic deviation threshold of the EC sensor, and using the dynamic deviation threshold to replace the deviation calibration threshold.

[0019] Preferably, the step of calculating the dynamic deviation threshold of the EC sensor and replacing the deviation calibration threshold with the dynamic deviation threshold specifically comprises the following steps:

[0020] Obtain the measurement data of the EC sensor in the past unit measurement cycle and calculate the average difference of the measurement values ​​in the past unit measurement cycle

[0021] The standard deviation SD of the measured value is calculated by a preset standard deviation formula, which is specifically:

[0022]

[0023] Where n is the total number of measurements in the measurement data of the EC sensor in the past unit measurement cycle, x i is the measurement value of the ith measurement in the measurement data of the EC sensor in the past unit measurement cycle;

[0024] The dynamic deviation threshold of the EC sensor is calculated as K×SD, where K is the dynamic calibration coefficient in the calibration threshold, and the dynamic threshold range of the measured value is determined as The deviation calibration threshold is replaced by a dynamic deviation threshold.

[0025] Preferably, the obtaining of the calibration request and matching and generating the calibration solution specifically include the following steps:

[0026] Obtain a calibration request, and generate multiple alternative solutions through matching with a preset calibration matching model, wherein 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 scheme score of each alternative scheme is calculated by a preset comprehensive score calculation formula; the comprehensive score calculation formula is specifically:

[0028]

[0029] Among them, O j is the cost of the standard calibration solution for the jth alternative, C j is the average number of times the standard calibration solution of the jth alternative is used, O is the preset calibration cost benchmark; P j is the calibration time of the jth alternative, P is the preset calibration time benchmark; Q j is the application environment temperature of the jth alternative, H j is the calibration environment temperature of the jth alternative, Q is the preset calibration temperature difference benchmark; Z1 is the cost coefficient, Z2 is the efficiency coefficient, Z3 is the environment coefficient, and Z1, Z2, and Z3 are all set by the management personnel;

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

[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, call or configure the calibration solution formula, and according to the calibration process information, start the EC sensor in the standard calibration environment to perform automatic calibration to obtain the calibration coefficient;

[0034] Input the calibration scheme and calibration coefficient into a preset regression validation model for validation; the regression validation model is a linear regression model obtained by iterative training of historical calibration data;

[0035] If the verification is passed, the automatic calibration is completed;

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

[0037] Preferably, the method further comprises: collecting measurement data and calibration process data of the EC sensor in real time, and regularly packaging and sending the data to the calibration matching model for negative feedback iterative training.

[0038] In a second aspect, the present application provides an EC sensor automatic calibration system, which adopts the following technical solution:

[0039] An EC sensor automatic calibration system, comprising:

[0040] A scheme matching module is used to obtain a calibration request and match and generate a calibration scheme; the calibration request includes basic information of the EC sensor, 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;

[0041] The measurement monitoring module is used to obtain the measurement data of the EC sensor in real time and to trigger the EC sensor based on the 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 time is met or the calibration trigger condition is met.

[0043] Preferably, the automatic calibration module comprises:

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

[0045] An automatic calibration unit is used to retrieve or configure a calibration solution formula according to a calibration scheme, and to start an EC sensor for automatic calibration in a standard calibration environment according to the calibration process information to obtain a calibration coefficient;

[0046] The calibration monitoring unit is used to input the calibration scheme and calibration coefficient into a preset regression validation model for validation; the regression validation model is a linear regression model obtained by iterative training of historical calibration data; if the validation passes, the automatic calibration is completed; if the validation fails, an error message is generated and sent to the management personnel.

[0047] Preferably, the solution matching module includes:

[0048] A scheme matching unit is used to obtain a calibration request and generate multiple alternative schemes by matching a preset calibration matching model, wherein 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 by using a preset comprehensive scoring calculation formula; the comprehensive scoring calculation formula is specifically:

[0050]

[0051] Among them, O j is the cost of the standard calibration solution for the jth alternative, C j is the average number of times the standard calibration solution of the jth alternative is used, O is the preset calibration cost benchmark; P j is the calibration time of the jth alternative, P is the preset calibration time benchmark; Q j is the application environment temperature of the jth alternative, H j is the calibration environment temperature of the jth alternative, Q is the preset calibration temperature difference benchmark; Z1 is the cost coefficient, Z2 is the efficiency coefficient, Z3 is the environment coefficient, and Z1, Z2, and Z3 are all set by the management personnel;

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

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

[0054] 1. Generate a calibration plan based on the calibration request through intelligent decision-making of the calibration matching model. It has strong universality and adapts to various application scenarios. It does not need to spend a lot of manpower and material resources on calibration preparation, and realizes accurate and efficient automatic calibration of the sensor. In the actual use of the EC sensor, it can judge whether the sensor is faulty based on the current measurement data and environmental conditions of the sensor. For EC sensors with abnormal data, it can directly calibrate to ensure the stable measurement of the EC sensor. When the EC sensor needs to be calibrated, the EC sensor is automatically calibrated according to the calibration plan, which reduces manual intervention, improves calibration consistency, and effectively improves the calibration efficiency and calibration accuracy of the EC sensor.

[0055] 2. Generate multiple alternative solutions that meet the calibration request based on the calibration request matching through the calibration matching model, and then score each alternative solution based on the calibration cost, solution preservation difficulty, calibration efficiency, and the difference between the calibration environment and the application environment, and then decide on the best calibration solution, so as to achieve more accurate calibration and improve 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 the calibration, which helps to remove the interference of irrelevant data, and then perform double verification of deviation and change rate. By first determining the deviation difference between the measured value and the historical average measured value, and then checking the change rate, a double verification mechanism is constructed, which can more comprehensively detect abnormal conditions in the measurement data, accurately locate the problem type, respond to the abnormality in a timely manner, and calibrate the EC sensor with abnormalities in advance to ensure the stable measurement of the EC sensor. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a method flow chart of an EC sensor automatic calibration method in an embodiment of the present application;

[0058] Figure 2 is a flow chart of a method for matching and generating a calibration scheme in an embodiment of the present application;

[0059] Figure 3 is a flow chart of a method for triggering and verifying an EC sensor in an embodiment of the present application;

[0060] Figure 4 is a flow chart of a method for calculating a dynamic deviation threshold of an EC sensor in an embodiment of the present application;

[0061] Figure 5 is a flow chart of a method for automatically calibrating an EC sensor in an embodiment of the present application;

[0062] Figure 6 It is a system block diagram of an EC sensor automatic calibration system in an embodiment of the present application.

[0063] Explanation of the accompanying drawings: 1. Scheme matching module; 11. Scheme matching unit; 12. Scheme scoring unit; 13. Scheme selection unit; 2. Measurement monitoring module; 3. Automatic calibration module; 31. Electrode cleaning unit; 32. Automatic calibration unit; 33. Calibration monitoring unit. DETAILED DESCRIPTION

[0064] The following is combined with Figure 1-6 This application is described in further detail.

[0065] The present application embodiment discloses an EC sensor automatic calibration method. Figure 1 , an EC sensor automatic calibration method, comprising the following steps:

[0066] S1. Obtain a calibration request and generate a calibration scheme by matching; the calibration request includes basic information of the EC sensor, 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;

[0067] S2, obtaining the measurement data of the EC sensor in real time, and performing trigger verification on the EC sensor based on the calibration trigger condition;

[0068] S3. When the calibration interval is met or the calibration trigger condition is met, the EC sensor is automatically calibrated based on the calibration scheme. Through the above steps, the calibration scheme is generated based on the calibration request through the intelligent decision-making of the calibration matching model. It has strong universality and adapts to a variety of application scenarios. There is no need to spend a lot of manpower and material resources on calibration preparation work, and the sensor is accurately and efficiently calibrated automatically. In the actual use of the EC sensor, it can be judged whether the sensor is faulty based on the current measurement data and environmental conditions of the sensor. For EC sensors with abnormal data, they can be calibrated directly to ensure the stable measurement of the EC sensor; and when the EC sensor needs to be calibrated, the EC sensor is automatically calibrated according to the calibration scheme to reduce manual intervention and improve calibration consistency, so as to effectively improve the calibration efficiency and calibration accuracy of the EC sensor.

[0069] Reference Figure 2 The obtaining of the calibration request and matching and generating the calibration scheme specifically include the following steps:

[0070] A1. Generate alternative solutions: Obtain a calibration request, and generate multiple alternative solutions by matching a preset 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 scheme score: Calculate the scheme score of each alternative scheme by using a preset comprehensive score calculation formula; the comprehensive score calculation formula is specifically:

[0072]

[0073] Among them, O j is the cost of the standard calibration solution for the jth alternative, C j is the average number of times the standard calibration solution of the jth alternative is used, O is the preset calibration cost benchmark; P j is the calibration time of the jth alternative, P is the preset calibration time benchmark; Q j is the application environment temperature of the jth alternative, H j is the calibration environment temperature of the jth alternative, Q is the preset calibration temperature difference benchmark; Z1 is the cost coefficient, Z2 is the efficiency coefficient, Z3 is the environment coefficient, and Z1, Z2, and Z3 are all set by the management personnel;

[0074] A3. Select the calibration scheme: Sort and compare the alternative schemes based on the scheme scores, and select the alternative scheme with the highest scheme score as the calibration scheme. The calibration matching model generates multiple alternative schemes that meet the calibration request based on the calibration request matching, and then scores each alternative scheme from the aspects of calibration cost (solution cost), solution preservation difficulty (the easier the solution is to preserve, the lower the single cost), calibration efficiency, and the difference between the calibration environment and the application environment (calibration convenience), and then decides on the best calibration scheme, so as to achieve more accurate calibration and improve measurement accuracy.

[0075] In addition, when the EC sensor is automatically calibrated, the measurement data and calibration process data of the EC sensor are collected in real time, and regularly packaged and sent to the calibration matching model for negative feedback iterative training. This helps to further improve and optimize the calibration matching model, simplify the calibration preparation process, and effectively improve the accuracy and advancement of the calibration scheme matching.

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

[0077] B1. Data acquisition and cleaning: acquiring the measurement data of the EC sensor in real time, and performing data cleaning on 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;

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

[0079] B3. If it does not exceed, the calibration trigger verification is determined to have passed;

[0080] B4. If it exceeds, the measurement deviation of the measurement item is determined, and then 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 specifically: change rate = (this measurement value - last measurement value) / last measurement value × 100%;

[0081] B5. Calibration trigger verification passed: If not exceeded, the calibration trigger verification is determined to have passed, and a sensor degradation prompt is generated and sent to the management staff;

[0082] B6. The calibration trigger condition is met: If it is exceeded, it is determined that the calibration trigger verification has not passed and the calibration trigger condition is met. By cleaning the data according to the calibration trigger condition, it is possible to focus on the measurement items that are truly related to the calibration, which helps to remove the interference of irrelevant data, and then perform a double check of the deviation and the rate of change. By first determining the deviation difference between the measured value and the historical average measured value, and then checking the rate of change, a double check mechanism is constructed, which can more comprehensively detect abnormalities in the measurement data, accurately locate the problem type, respond to abnormalities in a timely manner, and calibrate the EC sensor with abnormalities in advance to ensure the stable measurement of the EC sensor.

[0083] In addition, the generation of sensor degradation prompts to be sent to managers also includes: when managers read the sensor degradation prompts, a request to enable the dynamic deviation threshold is generated synchronously. After the manager agrees to start, the dynamic deviation threshold of the EC sensor is calculated, and the deviation calibration threshold is replaced by the dynamic deviation threshold. In the case where the calibration trigger validation passes but the measurement deviation is found, a sensor degradation prompt is generated and sent to the manager, which enables the manager to understand the working status of the sensor in a timely manner, and the dynamic deviation threshold is used for the degraded sensor. In order to save costs, when the sensor continues to be used, the monitoring capability is improved through adaptive adaptation, the sensor changes are dynamically adapted, the trigger accuracy is optimized, and unnecessary calibration operations are reduced to extend the service life of the sensor. At the same time, managers can make flexible decisions based on actual conditions. This flexibility enables the most suitable calibration trigger strategy to be selected when facing different sensors, different measurement tasks and different resource constraints. For example, for some chemical experiment monitoring scenarios with extremely high accuracy requirements, the dynamic deviation threshold may be enabled when the sensor shows signs of slight degradation to ensure the accuracy of the experimental data; while for some environmental monitoring scenarios with slightly lower accuracy requirements, it can be enabled after the sensor degrades to a certain extent to balance accuracy and cost.

[0084] Reference Figure 4 , calculating the dynamic deviation threshold of the EC sensor, and using the dynamic deviation threshold to replace the deviation calibration threshold specifically includes the following steps:

[0085] C1. Data collection: Obtain the measurement data of the EC sensor in the past unit measurement cycle and calculate the average difference of the measurement values ​​in the past unit measurement cycle

[0086] C2. Calculate the standard deviation of the measured value: Calculate the standard deviation SD of the measured value using a preset standard deviation formula. The standard deviation formula is specifically:

[0087]

[0088] Where n is the total number of measurements in the measurement data of the EC sensor in the past unit measurement cycle, x i is the measurement value of the ith measurement in 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 then determine the dynamic threshold range of the measured value as Use dynamic deviation threshold to replace deviation calibration threshold. When the sensor is found to be degraded, the management personnel decide to use dynamic deviation threshold to replace deviation calibration threshold through the above steps to ensure the accuracy of experimental data and realize intelligent automatic calibration.

[0090] Reference Figure 5 The automatic calibration of the EC sensor 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 perform automatic calibration according to the calibration process information to obtain the calibration coefficient;

[0093] D3. Calibration monitoring: input the calibration scheme and calibration coefficient into a preset regression validation model for validation; the regression validation model is a linear regression model obtained by iterative training of historical calibration data;

[0094] D4. If the verification is passed, the automatic calibration is completed;

[0095] D5. Generate an error: If the verification fails, an error message is generated and sent to the management staff. After completing the sensor calibration, a desired calibration coefficient range is predicted based on the calibration scheme through the regression verification model. If the actual calibration coefficient exceeds this range, there may be a problem, and an error message is generated and sent to the management staff. Through the above steps, the sensor calibration work can be monitored and evaluated in real time, potential calibration errors can be discovered in advance, and troubleshooting guidance can be provided for EC sensors with abnormal calibration, ensuring stable detection of EC sensors, helping to improve the measurement quality of EC sensors, and preventing erroneous data from being used in subsequent measurement processes, thereby effectively improving the correction efficiency and calibration accuracy of EC sensors.

[0096] Take an example to explain in detail. After the electrode is cleaned, the electrode of the EC sensor to be calibrated is placed in the standard calibration solution, and the EC sensor is started for measurement. In this example, the measurement item is the conductivity of the solution, and the EC sensor starts to measure the conductivity of the standard solution. This measurement process generally lasts for a period of time to obtain stable and reliable measurement data. During the measurement process, the sensor will collect data according to its own measurement principle (such as the potential difference measurement between electrodes, etc.); assuming that the measured conductivity value of the standard solution is EC1, and the known conductivity value of the standard solution is EC2, the calibration coefficient F = EC2 / EC1. This coefficient will be used to correct the subsequent actual measurement data. In the subsequent actual measurement, assuming that the uncorrected measurement data is EC3, the corrected measurement data EC4 = EC3*F. This is the existing EC sensor automatic calibration process. In the above steps, after calibration, the calibration scheme and calibration coefficient are input into the preset regression validation model. The regression validation model is a linear regression model in the machine learning model. The linear regression model can predict the calibration coefficient under normal conditions by fitting the linear relationship between the calibration coefficient and the standard solution concentration, temperature and other factors. After each calibration of the EC sensor, the new calibration coefficient and the corresponding standard solution information, environmental parameters and other data (calibration coefficient and calibration scheme) are input into the trained machine learning model. The model calculates and judges based on the input data. The model will predict an expected calibration coefficient range. If the actual calibration coefficient exceeds this range, there may be a problem, that is, the validation fails.

[0097] The present application also discloses an EC sensor automatic calibration system. Figure 6 , an EC sensor automatic calibration system, comprising:

[0098] Scheme matching module 1 is used to obtain a calibration request and match and generate a calibration scheme; the calibration request includes basic information of the EC sensor, 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;

[0099] The measurement monitoring module 2 is used to obtain the measurement data of the EC sensor in real time and to perform trigger verification on the EC sensor based on the calibration trigger condition;

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

[0101] In addition, it also includes a negative feedback module 4, which is used to collect the measurement data and calibration process data of the EC sensor in real time when the EC sensor is automatically calibrated, and regularly package and send them to the calibration matching model for negative feedback iterative training. According to the calibration request, the calibration matching model intelligently makes decisions to generate a calibration plan, which has strong universality and adapts to various application scenarios. There is no need to spend a lot of manpower and material resources on calibration preparation work, and the sensor is accurately and efficiently calibrated automatically. In the actual use of the EC sensor, it can be judged whether the sensor is faulty according to the current measurement data and environmental conditions of the sensor, and the EC sensor with abnormal data can be calibrated directly to ensure the stable measurement of the EC sensor; when the EC sensor needs to be calibrated, the EC sensor is automatically calibrated according to the calibration plan, reducing manual intervention, improving calibration consistency, and effectively improving the calibration efficiency and calibration accuracy of the EC sensor.

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

[0103] A scheme matching unit 11 is used to obtain a calibration request and generate multiple alternative schemes by matching a preset calibration matching model, wherein 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 alternative scheme by using a preset comprehensive scoring calculation formula; the comprehensive scoring calculation formula is specifically:

[0105]

[0106] Among them, O j is the cost of the standard calibration solution for the jth alternative, C jis the average number of times the standard calibration solution of the jth alternative is used, O is the preset calibration cost benchmark; P j is the calibration time of the jth alternative, P is the preset calibration time benchmark; Q j is the application environment temperature of the jth alternative, H j is the calibration environment temperature of the jth alternative, Q is the preset calibration temperature difference benchmark; Z1 is the cost coefficient, Z2 is the efficiency coefficient, Z3 is the environment coefficient, and Z1, Z2, and Z3 are all set by the management personnel;

[0107] The scheme selection unit 13 is used to sort and compare the alternative schemes based on the scheme score, and select the alternative scheme with the highest scheme score as the calibration scheme. Through the calibration matching model, multiple alternative schemes that meet the calibration request are generated based on the calibration request matching, and then the alternative schemes are scored from the aspects of calibration cost (solution cost), solution preservation difficulty (the easier the solution is to preserve, the lower the single cost), calibration efficiency, and the difference between the calibration environment and the application environment (calibration convenience), and then the best calibration scheme is decided, so as to achieve more accurate calibration and improve measurement accuracy.

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

[0109] An electrode cleaning unit 31, used for cleaning the electrodes of the EC sensor;

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

[0111] The calibration monitoring unit 33 is used to input the calibration scheme and calibration coefficient into a preset regression validation model for validation; the regression validation model is a linear regression model obtained by iterative training of historical calibration data; if the validation is passed, the automatic calibration is completed; if the validation is not passed, an error message is generated and sent to the management personnel. By cleaning the data according to the calibration trigger conditions, it is possible to focus on the measurement items that are truly related to the calibration, which helps to remove the interference of irrelevant data, and then perform a double check of the deviation and the rate of change. By first determining the deviation difference between the measured value and the historical average measured value, and then checking the rate of change, a double verification mechanism is constructed, which can more comprehensively detect abnormal conditions in the measurement data, accurately locate the type of problem, respond to the abnormality in a timely manner, and calibrate the EC sensor with abnormalities in advance to ensure the stable measurement of the EC sensor.

[0112] The embodiment of the present application also discloses a computer-readable storage medium, which stores a computer program that can be loaded by a processor and executed as in the above method. The computer-readable storage medium includes, for example: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0113] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on these embodiments, all other embodiments obtained by ordinary technicians in this field without making creative work 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, ordinary technicians in this field can still combine, add, delete or make other adjustments to the features in the various embodiments of the present invention according to the circumstances without conflict, without making creative work, so as to obtain different other technical solutions that do not deviate from the concept of the present invention in essence, and these technical solutions also belong to the scope of protection of the present invention.

Claims

1. An automatic calibration method for an EC sensor, characterized in that: The following steps are involved: Obtain a calibration request and match and generate a calibration scheme; the calibration request includes basic information of the EC sensor, 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; Acquire the measurement data of the EC sensor in real time and perform trigger verification on the EC sensor based on the calibration trigger conditions; When the calibration interval is met or the calibration trigger condition is met, the EC sensor is automatically calibrated based on the calibration scheme.

2. The EC sensor automatic calibration method according to claim 1, characterized in that: 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 include the following steps: Acquire measurement data of the EC sensor in real time, and perform data cleaning on the measurement data according to a calibration trigger condition to obtain a measurement value of a 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; Calculate the difference between the measured value of the measurement item and the historical average measured value of the EC sensor in the measurement cycle to obtain a deviation difference, and determine whether the deviation difference exceeds the deviation calibration threshold; If it does not exceed, the calibration trigger verification is determined to have passed; If it exceeds, the measurement deviation of the measurement item is determined, and then 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 specifically: change rate = (this measurement value - last measurement value) / last measurement value × 100%; If it does not exceed, the calibration trigger verification is determined to have passed, and a sensor degradation prompt is generated and sent to the management staff; If it exceeds, it is determined that the calibration trigger verification has not passed and the calibration trigger condition is met.

3. The EC sensor automatic calibration method according to claim 2, characterized in that: The generating of the sensor degradation prompt and sending it to the management personnel also includes: when the management personnel reads the sensor degradation prompt, synchronously generating a request to enable the dynamic deviation threshold, after the management personnel agrees to start, calculating the dynamic deviation threshold of the EC sensor, and using the dynamic deviation threshold to replace the deviation calibration threshold.

4. The EC sensor automatic calibration method according to claim 3, characterized in that: The step of calculating the dynamic deviation threshold of the EC sensor and replacing the deviation calibration threshold with the dynamic deviation threshold specifically comprises the following steps: Obtain the measurement data of the EC sensor in the past unit measurement cycle and calculate the average difference of the measurement values ​​in the past unit measurement cycle The standard deviation SD of the measured value is calculated by a preset standard deviation formula, which is specifically: Where n is the total number of measurements in the measurement data of the EC sensor in the past unit measurement cycle, x i is the measurement value of the ith measurement in the measurement data of the EC sensor in the past unit measurement cycle; The dynamic deviation threshold of the EC sensor is calculated as K×SD, where K is the dynamic calibration coefficient in the calibration threshold, and the dynamic threshold range of the measured value is determined as The deviation calibration threshold is replaced by a dynamic deviation threshold.

5. The EC sensor automatic calibration method according to claim 1, characterized in that: The obtaining of the calibration request and matching and generating the calibration solution specifically include the following steps: Obtain a calibration request, and generate multiple alternative solutions through matching with a preset calibration matching model, wherein 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; The scheme score of each alternative scheme is calculated by a preset comprehensive score calculation formula; the comprehensive score calculation formula is specifically: Among them, O j is the cost of the standard calibration solution for the jth alternative, C j is the average number of times the standard calibration solution of the jth alternative is used, O is the preset calibration cost benchmark; P j is the calibration time of the jth alternative, P is the preset calibration time benchmark; Q j is the application environment temperature of the jth alternative, H j is the calibration environment temperature of the jth alternative, Q is the preset calibration temperature difference benchmark; Z1 is the cost coefficient, Z2 is the efficiency coefficient, Z3 is the environment coefficient, and Z1, Z2, and Z3 are all set by the management personnel; The alternative plans are ranked and compared based on their plan scores, and the alternative plan with the highest plan score is selected as the calibration plan.

6. The EC sensor automatic calibration method according to claim 1, characterized in that: The automatic calibration of the EC sensor based on the calibration scheme specifically includes the following steps: Clean the electrodes of the EC sensor; According to the calibration scheme, call or configure the calibration solution formula, and according to the calibration process information, start the EC sensor in the standard calibration environment to perform automatic calibration to obtain the calibration coefficient; Input the calibration scheme and calibration coefficient into a preset regression validation model for validation; the regression validation 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 fails, an error message is generated and sent to the administrator.

7. The EC sensor automatic calibration method according to claim 5, characterized in that: Also includes: The measurement data and calibration process data of the EC sensor are collected in real time, and are regularly packaged and sent to the calibration matching model for negative feedback iterative training.

8. An EC sensor automatic calibration system, characterized in that: include: A scheme matching module (1) is used to obtain a calibration request and match and generate a calibration scheme; the calibration request includes basic information of the EC sensor, 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; A measurement monitoring module (2), used for acquiring measurement data of the EC sensor in real time and performing trigger verification on the EC sensor based on a calibration trigger condition; The automatic calibration module (3) is used to automatically calibrate the EC sensor based on the calibration scheme when the calibration interval time is met or the calibration trigger condition is met.

9. The EC sensor automatic calibration system according to claim 8, characterized in that: The automatic calibration module (3) comprises: An electrode cleaning unit (31), used for cleaning the electrodes of the EC sensor; An automatic calibration unit (32) is used to retrieve or configure a calibration solution formula according to a calibration scheme, and to start the EC sensor in a standard calibration environment to perform automatic calibration according to the calibration process information to obtain a calibration coefficient; The calibration monitoring unit (33) is used to input the calibration scheme and calibration coefficient into a preset regression validation model for validation; the regression validation model is a linear regression model obtained by iterative training of historical calibration data; if the validation passes, the automatic calibration is completed; if the validation fails, an error message is generated and sent to a management personnel.

10. The EC sensor automatic calibration system according to claim 8, characterized in that: The solution matching module (1) comprises: A scheme matching unit (11) is used to obtain a calibration request and generate multiple alternative schemes by matching a preset calibration matching model, wherein the calibration matching model is a convolutional neural network model obtained by deep learning using historical working data and calibration data of EC sensors in different application fields as samples; The scheme scoring unit (12) is used to calculate the scheme score of each alternative scheme by using a preset comprehensive scoring calculation formula; the comprehensive scoring calculation formula is specifically: Among them, O j is the cost of the standard calibration solution for the jth alternative, C j is the average number of times the standard calibration solution of the jth alternative is used, O is the preset calibration cost benchmark; P j is the calibration time of the jth alternative, P is the preset calibration time benchmark; Q j is the application environment temperature of the jth alternative, H j is the calibration environment temperature of the jth alternative, Q is the preset calibration temperature difference benchmark; Z1 is the cost coefficient, Z2 is the efficiency coefficient, Z3 is the environment coefficient, and Z1, Z2, and Z3 are all set by the management personnel; The scheme selection unit (13) is used to sort and compare the alternative schemes based on the scheme scores, and select the alternative scheme with the highest scheme score as the calibration scheme.

Images