Deep learning based electrochemical oxygen sensor temperature and pressure adaptive compensation method

Abstract

This invention discloses a deep learning-based adaptive compensation method for temperature and pressure in an electrochemical oxygen sensor, relating to the field of oxygen content detection technology. The method includes the following steps: constructing a sealed test space and simultaneously conducting random oxygen content tests within the sealed test space, recording the oxygen content test data; analyzing the temperature offset relationship between the ambient temperature and the sensor temperature based on the oxygen content test data; analyzing the pressure offset relationship between the ambient pressure and the sensor temperature based on the oxygen content test data; analyzing the primary and secondary offset relationships based on the temperature and pressure offset relationships; and adaptively compensating the sensor temperature based on the primary and secondary offset relationships. This invention addresses the problem that existing oxygen content detection technologies using electrochemical oxygen sensors cannot eliminate reading deviations under different temperature and pressure environments, resulting in inaccurate oxygen content measurements.

Description

Technical Field

[0001] This invention relates to the field of oxygen content detection technology, specifically to a deep learning-based adaptive compensation method for temperature and pressure in electrochemical oxygen sensors. Background Technology

[0002] Oxygen content detection technology refers to a series of methods, principles, and instrument systems that use physical, chemical, or biological methods to qualitatively identify and quantitatively measure the concentration or partial pressure of oxygen in gaseous or liquid media. Its core objective is to accurately, quickly, and reliably obtain the oxygen content, and it is one of the fundamental key technologies for environmental monitoring, industrial production, medical diagnosis, safety protection, and scientific research.

[0003] Existing oxygen content detection technologies typically generate oxygen ions through a reaction with oxygen, producing a current signal. This current signal is then measured to determine the oxygen content. However, the output signal of traditional electrochemical oxygen sensors is significantly affected by temperature and pressure. Under different temperature and pressure conditions, the readings will exhibit varying degrees of deviation. For example, patent application CN118937449A discloses an "electrochemical oxygen sensor," but this solution cannot eliminate reading deviations caused by temperature and pressure when detecting oxygen content, resulting in inaccurate oxygen content readings. Existing oxygen content detection technologies also suffer from the problem that electrochemical oxygen sensors cannot eliminate reading deviations under different temperature and pressure conditions, leading to inaccurate oxygen content measurements. Summary of the Invention

[0004] This invention aims to at least partially solve one of the technical problems in the prior art. By constructing a sealed test space and equipping it with experimental devices, random oxygen content tests are conducted within this sealed test space, and the oxygen content test data is recorded. Then, the temperature offset relationship between the ambient temperature and the sensor temperature is analyzed based on the oxygen content test data. Simultaneously, the pressure offset relationship between the ambient pressure and the sensor temperature is analyzed based on the oxygen content test data. Furthermore, the primary and secondary offset relationships are analyzed based on the temperature and pressure offset relationships. Next, the ambient temperature and pressure of the electrochemical oxygen sensor are obtained, and calibration reference points are extracted from the primary and secondary offset relationships based on the ambient temperature and pressure. Finally, adaptive compensation is performed on the real-time readings of the electrochemical oxygen sensor based on the calibration reference points. This addresses the problem that existing oxygen content detection technologies, when using electrochemical oxygen sensors to detect oxygen content, cannot eliminate reading deviations under different temperature and pressure environments, resulting in inaccurate oxygen content measurements.

[0005] To achieve the above objectives, this application provides a deep learning-based adaptive compensation method for temperature and pressure in electrochemical oxygen sensors, comprising the following steps:

[0006] Construct a closed test space and conduct random oxygen content tests within the closed test space, recording the oxygen content test data;

[0007] Analysis of the temperature deviation relationship between ambient temperature and sensor temperature based on oxygen content test data;

[0008] Analysis of the pressure offset relationship between environmental pressure and sensor temperature based on oxygen content test data;

[0009] Analyze the primary and secondary offset relationships based on temperature and pressure offset relationships;

[0010] The sensor temperature is adaptively compensated based on the main offset relationship and the auxiliary offset relationship.

[0011] Furthermore, a closed test space is constructed, and random oxygen content tests are conducted within this space. The oxygen content test data is recorded, including the following sub-steps:

[0012] Construct a sealed testing space and equip it with experimental devices;

[0013] Oxygen-containing random tests were conducted in a closed test space, and the oxygen-containing test data were recorded.

[0014] Furthermore, constructing a sealed test space and equipping it with experimental devices includes the following sub-steps:

[0015] A sealed test space is constructed, which does not exchange air with the outside. The sealed test space is equipped with an air inlet and an air outlet. The air outlet can discharge the gas in the sealed test space, making the sealed test space a vacuum state. The air inlet can input gas into the sealed test space.

[0016] The sealed test space is also equipped with a temperature control device and a pressure control device. The temperature control device can adjust the ambient temperature in the sealed test space, and the pressure control device can adjust the ambient pressure in the sealed test space.

[0017] Furthermore, conducting random oxygen content tests and recording oxygen content test data in a closed test space includes the following sub-steps:

[0018] In each random oxygen test, the ambient temperature and pressure are randomly configured for the closed test space.

[0019] An air with a known oxygen content is input into a sealed test space, and the known oxygen content is named the standard reading. The oxygen content in the sealed test space is detected by an electrochemical oxygen sensor, and the oxygen content output by the electrochemical oxygen sensor is named the deviation reading. The ambient temperature, ambient pressure, standard reading, and deviation reading together constitute the oxygen content test data.

[0020] Furthermore, analyzing the temperature deviation relationship between ambient temperature and sensor temperature based on oxygen content test data includes the following sub-steps:

[0021] Each oxygen test data point includes one each of ambient temperature, ambient pressure, standard reading, and deviation reading. Ambient temperature is labeled as HT, ambient pressure as HP, standard reading as BS, and deviation reading as PS.

[0022] Calculate PS-BS and name the calculation result as oxygen deviation, represented by the symbol OD.

[0023] Establish a two-dimensional coordinate system with HT as the X-axis and OD as the Y-axis, and name it Temperature Deviation Analysis Chart. Enter OD into the Temperature Deviation Analysis Chart according to HT.

[0024] A regression model is introduced, and the temperature deviation analysis graph is regressed using the regression model to output a temperature offset function. The temperature offset function and the temperature deviation analysis graph constitute the temperature offset relationship.

[0025] Furthermore, analyzing the pressure offset relationship between environmental pressure and sensor temperature based on oxygen content test data includes the following sub-steps:

[0026] Establish a two-dimensional coordinate system with HP as the X-axis and OD as the Y-axis, and name it Pressure Deviation Analysis Chart. Enter OD into the Pressure Deviation Analysis Chart according to HP.

[0027] The pressure deviation analysis chart is regressed using a regression model to output a pressure offset function. The pressure offset function and the pressure deviation analysis chart constitute the pressure offset relationship.

[0028] Furthermore, the analysis of the primary and secondary offset relationships based on temperature and pressure offset relationships includes the following sub-steps:

[0029] Obtain the R-squared values ​​of the temperature offset relationship and the pressure offset relationship, and name them temperature R-squared and pressure R-squared, respectively.

[0030] The temperature R-square and pressure R-square are compared. If the temperature R-square is greater than or equal to the pressure R-square, the temperature offset relationship is used as the primary offset relationship and the pressure offset relationship is used as the secondary offset relationship. If the temperature R-square is less than the pressure R-square, the pressure offset relationship is used as the primary offset relationship and the temperature offset relationship is used as the secondary offset relationship.

[0031] The temperature offset function and the pressure offset function are collectively referred to as the offset function to be analyzed, and the temperature deviation analysis chart and the pressure deviation analysis chart are collectively referred to as the offset analysis chart to be analyzed.

[0032] The function to be analyzed and the offset analysis graph to be analyzed in the main offset relation are named the main offset function and the main offset analysis graph, respectively. The function to be analyzed and the offset analysis graph to be analyzed in the auxiliary offset relation are named the auxiliary offset function and the auxiliary offset analysis graph, respectively.

[0033] Furthermore, the adaptive compensation of sensor temperature based on the primary offset relationship and the secondary offset relationship includes the following sub-steps:

[0034] The ambient temperature and pressure of the electrochemical oxygen sensor are obtained, and the calibration reference points in the main offset relationship and the auxiliary offset relationship are extracted based on the ambient temperature and pressure.

[0035] Adaptive compensation is performed on the real-time readings of the electrochemical oxygen sensor based on the calibration reference point.

[0036] Furthermore, obtaining the ambient temperature and pressure of the electrochemical oxygen sensor and extracting the calibration reference points from the main and auxiliary offset relationships based on the ambient temperature and pressure includes the following sub-steps:

[0037] Map the primary offset function onto the primary offset analysis graph to obtain the primary offset curve, and map the secondary offset function onto the secondary offset analysis graph to obtain the secondary offset curve;

[0038] The ambient temperature and ambient pressure of the electrochemical oxygen sensor are obtained and named as real-time temperature and real-time pressure, respectively. If the main offset relationship is temperature offset relationship, then real-time temperature is marked as the main compensation parameter and real-time pressure is marked as the auxiliary compensation parameter. Otherwise, real-time pressure is marked as the main compensation parameter and real-time temperature is marked as the auxiliary compensation parameter.

[0039] Find the point in the main offset curve where the X-axis equals the main compensation parameter and name it the main compensation reference point. Name the coordinate point in the main offset analysis graph where the X-axis equals the main compensation parameter the main compensation historical reference point.

[0040] Name the compensation auxiliary parameter corresponding to the main compensation historical reference point as the historical auxiliary parameter. Find the point in the auxiliary offset curve where the X-axis is equal to the compensation auxiliary parameter and name it as the auxiliary compensation benchmark reference point. Find the point in the auxiliary offset curve where the X-axis is equal to the historical auxiliary parameter and name it as the auxiliary compensation historical reference point.

[0041] The main compensation reference point, the main compensation historical reference point, the auxiliary compensation reference point, and the auxiliary compensation historical reference point are collectively referred to as calibration reference points.

[0042] Furthermore, adaptive compensation of the real-time readings of the electrochemical oxygen sensor based on the calibration reference point includes the following sub-steps:

[0043] Obtain the Y-axis values ​​of the auxiliary compensation benchmark reference point and the auxiliary compensation historical reference point, and label them as YF1 and YF2 respectively. At the same time, obtain the Y-axis values ​​of the main compensation benchmark reference point and the main compensation historical reference point, and label them as YC1 and YC2 respectively.

[0044] Assuming the adaptive compensation value is Q, determine whether there is a primary compensation historical reference point. If yes, output the first compensation signal; otherwise, output the second compensation signal.

[0045] If the first compensation signal is output, calculate (Q-YC2) / |YC2|=(YF1-YF2) / |YF2|, solve for Q. If there are different main compensation historical reference points, calculate Q corresponding to each main compensation historical reference point, and calculate the average value of Q to obtain the adaptive compensation value.

[0046] If the second compensation signal is output, the midpoint of the auxiliary offset curve on the X-axis is obtained and named the estimated compensation point. The Y-axis value of the estimated compensation point is marked as YF3. (Q-YC1) / |YC1|=(YF1-YF3) / |YF3| is calculated, and Q is solved to obtain the adaptive compensation value.

[0047] The real-time reading of the electrochemical oxygen sensor is obtained, and the real-time reading is added to the adaptive compensation value to obtain the compensated oxygen content.

[0048] The beneficial effects of this invention are as follows: By constructing a closed test space and equipping it with experimental devices, random oxygen content tests are conducted and oxygen content test data is recorded within the closed test space. Then, the temperature deviation relationship between ambient temperature and sensor temperature is analyzed based on the oxygen content test data. Simultaneously, the pressure deviation relationship between ambient pressure and sensor temperature is analyzed based on the oxygen content test data. Furthermore, the primary and secondary deviation relationships are analyzed based on the temperature and pressure deviation relationships. The advantage lies in the fact that the temperature and pressure deviation relationships reveal, to some extent, the influence of ambient temperature and ambient pressure on the deviation of oxygen content detection. At the same time, the influence of ambient temperature and ambient pressure on the oxygen content detection results is not equal. By analyzing the influence between the two, the primary and secondary deviation relationships are identified, providing a data basis for subsequent compensation and improving the accuracy and effectiveness of oxygen content detection.

[0049] This invention acquires the ambient temperature and pressure of the electrochemical oxygen sensor and extracts calibration reference points from the primary and secondary offset relationships based on these parameters. Finally, it adaptively compensates the real-time readings of the electrochemical oxygen sensor based on these calibration reference points. The advantage lies in the fact that the primary offset relationship has a greater impact on the readings, therefore it should be the primary focus for compensation analysis. Simultaneously, the secondary offset relationship is used as an aid during the compensation process to further calculate the oxygen content deviation, ultimately compensating for the electrochemical oxygen sensor readings. The compensated readings eliminate the reading deviations caused by ambient temperature and pressure, further improving the accuracy and effectiveness of oxygen content detection. Attached Figure Description

[0050] Figure 1 This is a flowchart of the steps of the method of the present invention;

[0051] Figure 2 This is a schematic diagram of the temperature deviation analysis graph of the present invention;

[0052] Figure 3 This is a schematic diagram of the pressure deviation analysis diagram of the present invention;

[0053] Figure 4 This is a schematic diagram of the main offset curve of the present invention;

[0054] Figure 5 This is a schematic diagram of the auxiliary offset curve of the present invention;

[0055] Figure 6 This is a schematic diagram of the main compensation reference point and the main compensation historical reference point of the present invention.

[0056] Figure 7 This is a schematic diagram of the auxiliary compensation benchmark reference point and the auxiliary compensation historical reference point of the present invention. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Example 1, please refer to Figure 1 As shown, this application provides a deep learning-based adaptive compensation method for temperature and pressure in electrochemical oxygen sensors, comprising the following steps:

[0059] Step S1 involves constructing a sealed test space and conducting random oxygen content tests within that space, recording the oxygen content test data. Step S1 includes the following sub-steps:

[0060] Step S101: Construct a sealed test space and equip it with experimental devices;

[0061] Step S101 includes the following sub-steps:

[0062] Step S1011: Construct a sealed test space. The sealed test space does not exchange air with the outside. The sealed test space is equipped with an air inlet and an air outlet. The air outlet can discharge the gas in the sealed test space, making the sealed test space a vacuum state. The air inlet can input gas into the sealed test space.

[0063] Step S1012: A temperature control device and a pressure control device are also installed in the sealed test space. The temperature control device can adjust the ambient temperature in the sealed test space, and the pressure control device can adjust the ambient pressure in the sealed test space.

[0064] In practice, all devices used in the sealed test space are existing equipment that can be purchased directly. During installation, they can be directly installed inside the sealed test space without any specific restrictions on the installation location. Before each random oxygen test, the exhaust port will discharge all the gas in the sealed test space, making the sealed test space a vacuum state to prevent residual air inside from affecting the oxygen content of the subsequent gas.

[0065] Step S102: Conduct random oxygen content tests in a closed test space and record the oxygen content test data;

[0066] Step S102 includes the following sub-steps:

[0067] Step S1021: In each random oxygen test, the ambient temperature and ambient pressure are randomly configured for the closed test space;

[0068] Step S1022: Input air with a known oxygen content into the sealed test space, name the known oxygen content as the standard reading, detect the oxygen content in the sealed test space through an electrochemical oxygen sensor, name the oxygen content output by the electrochemical oxygen sensor as the deviation reading, and the ambient temperature, ambient pressure, standard reading and deviation reading together constitute the oxygen content test data.

[0069] In practice, after the sealed test space becomes a vacuum, air with a known oxygen content is introduced through the air inlet. For example, if air with an oxygen content of 21.45% is introduced into the sealed test space, the standard reading will be 21.45%, and the reading of the electrochemical oxygen sensor in the sealed test space will be 22.52%, which is the deviation reading. At this time, the randomly assigned ambient temperature is 32℃ and the ambient pressure is 80kPa. At this time, one oxygen content test data can be obtained. At least 1000 tests are performed to ensure that there is a sufficient sample size, that is, at least 1000 oxygen content test data.

[0070] Step S2 involves analyzing the temperature shift relationship between the ambient temperature and the sensor temperature based on the oxygen content test data. Step S2 includes the following sub-steps:

[0071] Step S201: Each oxygen test data includes one ambient temperature, one ambient pressure, one standard reading, and one deviation reading. The ambient temperature is marked as HT, the ambient pressure as HP, the standard reading as BS, and the deviation reading as PS.

[0072] Step S202: Calculate PS-BS and name the calculation result as oxygen deviation, represented by the symbol OD;

[0073] Please see Figure 2 As shown, in step S203, a two-dimensional coordinate system is established with HT as the X-axis and OD as the Y-axis, named the temperature deviation analysis chart, and OD is entered into the temperature deviation analysis chart according to HT.

[0074] Step S204: Introduce a regression model, regress the temperature deviation analysis graph using the regression model, and output the temperature offset function. The temperature offset function and the temperature deviation analysis graph are the temperature offset relationship.

[0075] In specific implementation, taking the oxygen content test data listed in step S101 as an example, where PS is 22.52% and BS is 21.45%, the calculated oxygen content deviation OD is 1.07%. At this time, HT is 32℃ and HP is 80kPa. A temperature deviation analysis chart is constructed, and the coordinate point (32℃, 1.07%) is entered into the temperature deviation analysis chart. All OD values ​​are entered into the temperature deviation analysis chart according to HT, resulting in the temperature deviation analysis chart as shown below. Figure 2 As shown, the temperature deviation analysis chart was regressed using a regression model, yielding the temperature offset function OD = 0.0002 × HT. 2 +0.091×HT-2.0146.

[0076] Step S3 involves analyzing the pressure offset relationship between ambient pressure and sensor temperature based on oxygen content test data. Step S3 includes the following sub-steps:

[0077] Please see Figure 3 As shown, in step S301, a two-dimensional coordinate system is established with HP as the X-axis and OD as the Y-axis, named the pressure deviation analysis chart, and OD is entered into the pressure deviation analysis chart according to HP.

[0078] Step S302: Regress the pressure deviation analysis diagram using a regression model to output the pressure offset function. The pressure offset function and the pressure deviation analysis diagram constitute the pressure offset relationship.

[0079] In practice, similar to step S2, the pressure deviation analysis diagram is constructed as follows: Figure 3 As shown, the pressure deviation analysis chart was regressed using a regression model, yielding the pressure offset function OD = -0.0001 × HP. 2 +0.0863×HP 2 -7.126.

[0080] Step S4 involves analyzing the primary and secondary offset relationships based on the temperature and pressure offset relationships. Step S4 includes the following sub-steps:

[0081] Step S401: Obtain the R-squared values ​​of the temperature offset relationship and the pressure offset relationship, and name them as temperature R-squared and pressure R-squared, respectively.

[0082] Step S402: Compare the temperature R square and the pressure R square. If the temperature R square is greater than or equal to the pressure R square, then the temperature offset relationship is taken as the main offset relationship and the pressure offset relationship is taken as the auxiliary offset relationship; if the temperature R square is less than the pressure R square, then the pressure offset relationship is taken as the main offset relationship and the temperature offset relationship is taken as the auxiliary offset relationship.

[0083] Step S403: The temperature offset function and the pressure offset function are collectively referred to as the offset function to be analyzed, and the temperature deviation analysis chart and the pressure deviation analysis chart are collectively referred to as the offset analysis chart to be analyzed.

[0084] Step S404: Name the function to be analyzed and the offset analysis graph to be analyzed in the main offset relationship as the main offset function and the main offset analysis graph, respectively; name the function to be analyzed and the offset analysis graph to be analyzed in the auxiliary offset relationship as the auxiliary offset function and the auxiliary offset analysis graph, respectively.

[0085] In practice, Rsquared is the Rregression of the regression function. 2The obtained temperature R-squared value is 0.8155, and the pressure R-squared value is 0.5349. By comparison, it was found that the temperature R-squared value is greater than the pressure R-squared value. Therefore, the temperature offset relationship is taken as the primary offset relationship, and the pressure offset relationship is taken as the secondary offset relationship. The temperature offset function is the primary offset function, and the pressure offset function is the secondary offset function. The temperature offset analysis plot is the primary offset analysis plot, and the pressure offset analysis plot is the secondary offset analysis plot. Since the R-squared value of the primary offset relationship is greater than that of the secondary offset relationship, it means that the primary offset relationship has a higher goodness of fit. Under different ambient temperature and ambient pressure conditions, the OD is more significantly and accurately affected by the primary offset relationship, while the influence of the secondary offset relationship is weaker.

[0086] Step S5 involves adaptively compensating for the sensor temperature based on the primary offset relationship and the secondary offset relationship. Step S5 includes the following sub-steps:

[0087] Step S501: Obtain the ambient temperature and ambient pressure of the electrochemical oxygen sensor and extract the calibration reference point in the main offset relationship and the auxiliary offset relationship based on the ambient temperature and ambient pressure.

[0088] Step S501 includes the following sub-steps:

[0089] Please see Figures 4 to 5 As shown, in step S5011, the main offset function is mapped onto the main offset analysis graph to obtain the main offset curve, and the auxiliary offset function is mapped onto the auxiliary offset analysis graph to obtain the auxiliary offset curve.

[0090] Step S5012: Obtain the ambient temperature and ambient pressure of the electrochemical oxygen sensor, and name them as real-time temperature and real-time pressure respectively. If the main offset relationship is temperature offset relationship, then mark the real-time temperature as the main compensation parameter and the real-time pressure as the auxiliary compensation parameter. Otherwise, mark the real-time pressure as the main compensation parameter and the real-time temperature as the auxiliary compensation parameter.

[0091] Please see Figure 6 As shown, in step S5013, find the point in the main offset curve where the X-axis is equal to the main compensation parameter, name it as the main compensation reference point, and name the coordinate point in the main offset analysis graph where the X-axis is equal to the main compensation parameter as the main compensation historical reference point.

[0092] Please see Figure 7 As shown, in step S5014, the compensation auxiliary parameter corresponding to the main compensation historical reference point is named as the historical auxiliary parameter. The point in the auxiliary offset curve whose X-axis is equal to the compensation auxiliary parameter is named as the auxiliary compensation benchmark reference point. The point in the auxiliary offset curve whose X-axis is equal to the historical auxiliary parameter is named as the auxiliary compensation historical reference point.

[0093] Step S5015: The main compensation reference point, the main compensation historical reference point, the auxiliary compensation reference point, and the auxiliary compensation historical reference point are collectively referred to as calibration reference points.

[0094] In practice, the mapping yields the main offset curve and the auxiliary offset curve as follows: Figure 4 and Figure 5 As shown, the real-time temperature is 26℃ and the real-time pressure is 80kPa. Since the main offset relationship is a temperature offset relationship, the main compensation parameter is the real-time temperature, and the auxiliary compensation parameter is the real-time pressure. The point on the main offset curve where the X-axis equals 26℃ is found, and the coordinates of the point on the main offset analysis graph where the X-axis equals 26℃ are also obtained. This yields the main compensation reference point and the main compensation historical reference point, as shown below. Figure 6 As shown, the coordinates of a certain primary compensation historical reference point are (26℃, 90kPa). The corresponding historical auxiliary parameter for this primary compensation historical reference point is 90kPa. Find the points in the auxiliary offset curve where the X-axis equals 80kPa, and simultaneously find the points in the auxiliary offset curve where the X-axis equals 90kPa. This yields the auxiliary compensation benchmark reference point and the auxiliary compensation historical reference point, as shown below. Figure 7 As shown.

[0095] Step S502: Adaptive compensation is performed on the real-time readings of the electrochemical oxygen sensor based on the calibration reference point;

[0096] Step S502 includes the following sub-steps:

[0097] Step S5021: Obtain the Y-axis values ​​of the auxiliary compensation reference point and the auxiliary compensation historical reference point, and label them as YF1 and YF2 respectively. At the same time, obtain the Y-axis values ​​of the main compensation reference point and the main compensation historical reference point, and label them as YC1 and YC2 respectively.

[0098] Step S5022: Assuming the adaptive compensation value is Q, determine whether there is a main compensation historical reference point. If yes, output the first compensation signal; otherwise, output the second compensation signal.

[0099] Step S5023: If the first compensation signal is output, calculate (Q-YC2) / |YC2|=(YF1-YF2) / |YF2|, solve for Q, if there are different main compensation historical reference points, calculate Q corresponding to each main compensation historical reference point, and calculate the average value of Q to obtain the adaptive compensation value;

[0100] In specific implementation, Figure 6 and Figure 7Taking the main compensation benchmark reference point, main compensation historical reference point, auxiliary compensation benchmark reference point, and auxiliary compensation historical reference point as an example, YC1 is obtained as 0.46%, YC2 as 1.32%, YF1 as -0.92%, and YF2 as -0.28%. Since a main compensation historical reference point exists, the first compensation signal is output, and (Q-YC2) / |YC2|=(YF1-YF2) / |YF2| is calculated, which is (Q-1.32%) / |1.32%|=(-0.92%+0.28%) / |-0.28%. The solution yields Q as -1.70%, with all results rounded to two decimal places. This indicates the adaptive compensation value is -1.70%. The calculation (Q-YC2) / |YC2|=(YF1-YF2) / |YF2| is based on the fact that (Q-YC2) / |YC2| represents the positional relationship between the current OD in the primary offset relationship and the OD in the historical data, while (YF1-YF2) / |YF2| represents the positional relationship between the current OD in the secondary offset relationship and the OD in the historical data. There is a relationship between the coordinates of the primary offset relationship and the primary offset curve. The fluctuation is caused by the influence of the auxiliary offset relationship. If the compensation master parameter and compensation auxiliary parameter are the same in both sets of data, then their oxygen content deviations should be very close or equal. When the compensation master parameter is the same, the biggest factor affecting the oxygen content deviation is the compensation auxiliary parameter, and the deviation distance of the coordinate points reveals the positional relationship between different compensation auxiliary parameters. Therefore, the oxygen content deviation under the current annular temperature and environmental pressure can be obtained by solving (Q-YC2) / |YC2|=(YF1-YF2) / |YF2|. In this example, (Q-YC2) can calculate the difference between the OD in the current environment and the OD in the historical data. If (Q-YC2) is negative, it means that the OD in the current environment should be smaller than the OD in the historical data. In (Q-YC2) / |YC2|, if no absolute value sign is added, when YC2 is negative, the calculation result will be positive. In this case, the OD in the current environment will be larger than the OD in the historical data. Therefore, an absolute value sign needs to be added. The same applies to (YF1-YF2) / |YF2|. In this example, no further explanation will be given.

[0101] Step S5024: If the second compensation signal is output, the midpoint of the auxiliary offset curve on the X-axis dimension is obtained and named the estimated compensation point. The value of the Y-axis of the estimated compensation point is marked as YF3. (Q-YC1) / |YC1|=(YF1-YF3) / |YF3| is calculated, and Q is solved to obtain the adaptive compensation value.

[0102] Step S5025: Obtain the real-time reading of the electrochemical oxygen sensor, add the real-time reading to the adaptive compensation value, and obtain the compensated oxygen content.

[0103] In specific implementation, if a second compensation signal is output, it indicates that there is no historical data for reference. In this case, the main offset curve and the auxiliary offset curve are used as references. Since the main offset curve and the auxiliary offset curve are obtained from the analysis of numerous historical data, their curves should approach the median value for each value of X. Therefore, the compensation auxiliary parameter corresponding to the main compensation reference point is regarded as the midpoint of the auxiliary offset curve in the X-axis dimension, i.e., the estimated compensation point. In this embodiment, the range of the auxiliary offset curve in the X-axis dimension is 50 kPa to 110 kPa, thus the estimated compensation point is 80 kPa, (Q-YC1) / |YC1| =(YF1-YF3) / |YF3| is actually calculating the positional relationship between the OD under the current environmental conditions and the main compensation reference point. The positional relationship is close to the positional relationship between the OD on the auxiliary offset curve and the estimated compensation point under the current environmental conditions. The calculation method and principle are the same as the first compensation signal, and will not be described in this embodiment. The calculated adaptive compensation value is -1.70%, which means that the reading of the electrochemical oxygen sensor is 1.7% smaller than the actual value. The real-time reading is 19.6%. The calculated oxygen content after compensation is 21.3% (19.6%-(-1.7%)).

[0104] Example 2: This application provides an electronic device, which may include a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The memory stores computer-readable instructions, and the processor can call these instructions. When the processor executes a computer-readable instruction, it performs steps such as those in the deep learning-based adaptive compensation method for temperature and pressure in an electrochemical oxygen sensor to achieve the following functions: constructing a sealed test space and simultaneously conducting random oxygen-containing tests within the sealed test space, recording the oxygen-containing test data; analyzing the temperature offset relationship between the ambient temperature and the sensor temperature based on the oxygen-containing test data; analyzing the pressure offset relationship between the ambient pressure and the sensor temperature based on the oxygen-containing test data; analyzing the primary and secondary offset relationships based on the temperature and pressure offset relationships; and adaptively compensating the sensor temperature based on the primary and secondary offset relationships.

[0105] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0106] Example 3: This application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the deep learning-based adaptive compensation method for temperature and pressure of an electrochemical oxygen sensor provided by the above methods. This method includes: constructing a sealed test space and simultaneously conducting random oxygen-containing tests within the sealed test space, recording oxygen-containing test data; analyzing the temperature offset relationship between the ambient temperature and the sensor temperature based on the oxygen-containing test data; analyzing the pressure offset relationship between the ambient pressure and the sensor temperature based on the oxygen-containing test data; analyzing the primary offset relationship and the secondary offset relationship based on the temperature offset relationship and the pressure offset relationship; and adaptively compensating the sensor temperature based on the primary offset relationship and the secondary offset relationship.

[0107] Example 4: This application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it performs the steps of the above-described deep learning-based adaptive compensation method for temperature and pressure of an electrochemical oxygen sensor to achieve the following functions: constructing a sealed test space and simultaneously conducting random oxygen-containing tests within the sealed test space, recording oxygen-containing test data; analyzing the temperature offset relationship between ambient temperature and sensor temperature based on the oxygen-containing test data; analyzing the pressure offset relationship between ambient pressure and sensor temperature based on the oxygen-containing test data; analyzing the primary offset relationship and secondary offset relationship based on the temperature offset relationship and pressure offset relationship; and adaptively compensating the sensor temperature based on the primary offset relationship and secondary offset relationship.

[0108] Based on the above description of the embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the above technical solutions, in essence or in terms of their contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments.

[0109] In the embodiments provided in this application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interfaces. The indirect coupling or communication connection between systems, modules, and units may be electrical, mechanical, or other forms.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A deep learning-based adaptive compensation method for temperature and pressure in electrochemical oxygen sensors, characterized in that, Includes the following steps: Construct a closed test space and conduct random oxygen content tests within the closed test space, recording the oxygen content test data; Analysis of the temperature deviation relationship between ambient temperature and sensor temperature based on oxygen content test data; Analysis of the pressure offset relationship between environmental pressure and sensor temperature based on oxygen content test data; Analyze the primary and secondary offset relationships based on temperature and pressure offset relationships; The sensor temperature is adaptively compensated based on the main offset relationship and the auxiliary offset relationship.

2. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 1, characterized in that, Constructing a closed test space and conducting random oxygen content tests within it, recording the oxygen content test data includes the following sub-steps: Construct a sealed testing space and equip it with experimental devices; Oxygen-containing random tests were conducted in a closed test space, and the oxygen-containing test data were recorded.

3. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 2, characterized in that, Setting up a sealed test space and equipping it with experimental devices includes the following sub-steps: A sealed test space is constructed, which does not exchange air with the outside. The sealed test space is equipped with an air inlet and an air outlet. The air outlet can discharge the gas in the sealed test space, making the sealed test space a vacuum state. The air inlet can input gas into the sealed test space. The sealed test space is also equipped with a temperature control device and a pressure control device. The temperature control device can adjust the ambient temperature in the sealed test space, and the pressure control device can adjust the ambient pressure in the sealed test space.

4. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 3, characterized in that, The process of conducting random oxygen content tests and recording oxygen content test data in a closed test space includes the following sub-steps: In each random oxygen test, the ambient temperature and pressure are randomly configured for the closed test space. An air with a known oxygen content is input into a sealed test space, and the known oxygen content is named the standard reading. The oxygen content in the sealed test space is detected by an electrochemical oxygen sensor, and the oxygen content output by the electrochemical oxygen sensor is named the deviation reading. The ambient temperature, ambient pressure, standard reading, and deviation reading together constitute the oxygen content test data.

5. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 4, characterized in that, Analyzing the temperature deviation relationship between ambient temperature and sensor temperature based on oxygen content test data includes the following sub-steps: Each oxygen test data point includes one each of ambient temperature, ambient pressure, standard reading, and deviation reading. Ambient temperature is labeled as HT, ambient pressure as HP, standard reading as BS, and deviation reading as PS. Calculate PS-BS and name the calculation result as oxygen deviation, represented by the symbol OD. Establish a two-dimensional coordinate system with HT as the X-axis and OD as the Y-axis, and name it Temperature Deviation Analysis Chart. Enter OD into the Temperature Deviation Analysis Chart according to HT. A regression model is introduced, and the temperature deviation analysis graph is regressed using the regression model to output a temperature offset function. The temperature offset function and the temperature deviation analysis graph constitute the temperature offset relationship.

6. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 5, characterized in that, Analyzing the pressure offset relationship between ambient pressure and sensor temperature based on oxygen content test data includes the following sub-steps: Establish a two-dimensional coordinate system with HP as the X-axis and OD as the Y-axis, and name it Pressure Deviation Analysis Chart. Enter OD into the Pressure Deviation Analysis Chart according to HP. The pressure deviation analysis chart is regressed using a regression model to output a pressure offset function. The pressure offset function and the pressure deviation analysis chart constitute the pressure offset relationship.

7. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 6, characterized in that, The analysis of the primary and secondary offset relationships based on temperature and pressure offset relationships includes the following sub-steps: Obtain the R-squared values ​​of the temperature offset relationship and the pressure offset relationship, and name them temperature R-squared and pressure R-squared, respectively. The temperature R-square and pressure R-square are compared. If the temperature R-square is greater than or equal to the pressure R-square, the temperature offset relationship is used as the primary offset relationship and the pressure offset relationship is used as the secondary offset relationship. If the temperature R-square is less than the pressure R-square, the pressure offset relationship is used as the primary offset relationship and the temperature offset relationship is used as the secondary offset relationship. The temperature offset function and the pressure offset function are collectively referred to as the offset function to be analyzed, and the temperature deviation analysis chart and the pressure deviation analysis chart are collectively referred to as the offset analysis chart to be analyzed. The function to be analyzed and the offset analysis graph to be analyzed in the main offset relation are named the main offset function and the main offset analysis graph, respectively. The function to be analyzed and the offset analysis graph to be analyzed in the auxiliary offset relation are named the auxiliary offset function and the auxiliary offset analysis graph, respectively.

8. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 7, characterized in that, Adaptive compensation of sensor temperature based on the primary and secondary offset relationships includes the following sub-steps: The ambient temperature and pressure of the electrochemical oxygen sensor are obtained, and the calibration reference points in the main offset relationship and the auxiliary offset relationship are extracted based on the ambient temperature and pressure. Adaptive compensation is performed on the real-time readings of the electrochemical oxygen sensor based on the calibration reference point.

9. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 8, characterized in that, The process of obtaining the ambient temperature and pressure of the electrochemical oxygen sensor and extracting calibration reference points from the main and auxiliary offset relationships based on the ambient temperature and pressure includes the following sub-steps: Map the primary offset function onto the primary offset analysis graph to obtain the primary offset curve, and map the secondary offset function onto the secondary offset analysis graph to obtain the secondary offset curve; The ambient temperature and ambient pressure of the electrochemical oxygen sensor are obtained and named as real-time temperature and real-time pressure, respectively. If the main offset relationship is temperature offset relationship, then real-time temperature is marked as the main compensation parameter and real-time pressure is marked as the auxiliary compensation parameter. Otherwise, real-time pressure is marked as the main compensation parameter and real-time temperature is marked as the auxiliary compensation parameter. Find the point in the main offset curve where the X-axis equals the main compensation parameter and name it the main compensation reference point. Name the coordinate point in the main offset analysis graph where the X-axis equals the main compensation parameter the main compensation historical reference point. Name the compensation auxiliary parameter corresponding to the main compensation historical reference point as the historical auxiliary parameter. Find the point in the auxiliary offset curve where the X-axis is equal to the compensation auxiliary parameter and name it as the auxiliary compensation benchmark reference point. Find the point in the auxiliary offset curve where the X-axis is equal to the historical auxiliary parameter and name it as the auxiliary compensation historical reference point. The main compensation reference point, the main compensation historical reference point, the auxiliary compensation reference point, and the auxiliary compensation historical reference point are collectively referred to as calibration reference points.

10. The temperature and pressure adaptive compensation method for an electrochemical oxygen sensor based on deep learning according to claim 9, characterized in that, Adaptive compensation of real-time readings of the electrochemical oxygen sensor based on a calibration reference point includes the following sub-steps: Obtain the Y-axis values ​​of the auxiliary compensation benchmark reference point and the auxiliary compensation historical reference point, and label them as YF1 and YF2 respectively. At the same time, obtain the Y-axis values ​​of the main compensation benchmark reference point and the main compensation historical reference point, and label them as YC1 and YC2 respectively. Assuming the adaptive compensation value is Q, determine whether there is a primary compensation historical reference point. If yes, output the first compensation signal; otherwise, output the second compensation signal. If the first compensation signal is output, calculate (Q-YC2) / |YC2|=(YF1-YF2) / |YF2|, solve for Q. If there are different main compensation historical reference points, calculate Q corresponding to each main compensation historical reference point, and calculate the average value of Q to obtain the adaptive compensation value. If the second compensation signal is output, the midpoint of the auxiliary offset curve on the X-axis is obtained and named the estimated compensation point. The Y-axis value of the estimated compensation point is marked as YF3. (Q-YC1) / |YC1|=(YF1-YF3) / |YF3| is calculated, and Q is solved to obtain the adaptive compensation value. The real-time reading of the electrochemical oxygen sensor is obtained, and the real-time reading is added to the adaptive compensation value to obtain the compensated oxygen content.

Images