A Tin Dioxide-Type Sensor and Method for Quantifying H2 and CO Gases

Abstract

The present invention relates to the field of metal oxide semiconductor gas sensors, and provides a tin dioxide type sensor and method for quantitatively detecting H2 and CO gases. The sensor includes a heating element, a gas sensitive element, a thermosensitive element, and a microprocessor. The heating element is a flat structure with a metal ceramic inner layer and an alumina outer cover. The gas sensitive element is a thin film formed by screen printing gas sensitive material nano-powder and then firing at high temperature on one side of the heating element. The thermosensitive element is on the other side of the heating element, feeding back the temperature signal of the gas sensitive element. The microprocessor controls the temperature change of the heating element, receives the signals of the gas sensitive element and the thermosensitive element, and calculates the type and concentration value of the gas according to the gas recognition non-linear model. In the present invention, the gas sensitive element, the heating element, and the thermistor are integrated to meet the fast response of the response signal to temperature changes; the gases are recognized by using the difference in gas adsorption; through signal acquisition, conversion and combination with a liquid crystal screen, the detection result is given quickly and intuitively.

Description

Technical Field

[0001] The present invention relates to the field of metal oxide semiconductor gas sensors, and particularly to a tin dioxide type sensor and method for quantitatively detecting H2 and CO gases. Background Art

[0002] Compared with other sensors, the SnO2 gas sensor has the advantages of simple structure, high sensitivity, low manufacturing cost and long service life, and is widely used in the detection of various flammable, explosive and harmful gas concentrations, and its usage is increasing year by year.

[0003] Working principle: The presence and concentration of the gas to be detected are judged by the release or acquisition of electrons due to the interaction between the gas to be detected and the gas adsorbed on the surface of the gas-sensitive material, which causes a change in the conductivity of the material. Theoretically speaking, the amount of electron gain and loss can be inferred from the type and number of adsorbed gas molecules. For example:

[0004] H2 + O 2- = H2O + 2e (1)

[0005] CO + O 2- = CO2 + 2e (2)

[0006] However, it is impossible to identify the gas type and quantitatively determine the concentration solely based on the change in the conductivity of the gas-sensitive material.

[0007] Existing solutions include:

[0008] (1) By loading and doping in the SnO2 material, the adsorption performance for specific gases is improved to enhance the selectivity for specific gases. However, for strong reducing gases such as H2 and CO, although the response value of H2 is nearly 10 times larger than that of CO at low concentrations after modification, it is very difficult to identify and eliminate the influence, especially in the case of large concentration changes.

[0009] (2) In recent years, some people have tried to identify gases based on the differences in gas adsorption and desorption on the material surface at different temperatures through temperature modulation, and there have been some successful reports. However, accurately controlling the temperature change of the gas-sensitive element in the latter is a difficult point, and there has been no application report yet. Summary of the Invention

[0010] The purpose of the present invention is to at least overcome one of the deficiencies of the prior art, and provide a tin dioxide type sensor and method for quantitatively detecting H2 and CO gases. According to the adsorption and desorption laws of gases on the surface of SnO2 and its modified materials at different temperatures and during the temperature change process, the gas type is identified and its content is quantitatively determined. Four key problems are mainly solved:

[0011] (1) The SnO2 thin film is fabricated by screen printing on a thin-film heating element, solving the problems of rapid temperature rise, temperature drop, and temperature followability of the gas sensor element.

[0012] (2) Through the PID temperature control technology, a rapid, stable temperature rise - constant temperature - temperature drop process is achieved, solving the interference problem of environmental changes on the gas sensor element.

[0013] (3) By optimizing the configuration of the signal acquisition and control module, rapid data transmission, conversion, and display are realized, solving the problem of signal distortion.

[0014] (4) Combining with the BP artificial neural network, a non-linear parameter matrix of resistance and gas concentration with respect to temperature change is optimized, solving the problems of gas recognition and quantification.

[0015] The present invention adopts the following technical solutions:

[0016] On the one hand, the present invention provides a tin dioxide type sensor for quantitatively detecting H2 and CO gases, comprising a heating element, a gas sensor element, a thermosensitive element, and a microprocessor;

[0017] The heating element is a flat structure with a cermet inner layer and an alumina outer cover, and the positive and negative electrodes are led out as the heating control terminals;

[0018] The gas sensor element is a thin film formed by screen printing a gas-sensitive material nano-powder and then firing it at a high temperature on one side of the heating element; interdigital electrodes are arranged between the gas sensor element and the heating element, and the positive and negative measurement terminals of the gas sensor element are led out;

[0019] The thermosensitive element is arranged on the other side of the heating element to feedback the temperature signal of the gas sensor element;

[0020] The microprocessor controls the temperature change of the heating element, receives the signals of the gas sensor element and the thermosensitive element, and calculates the type and concentration value of the detected gas according to the gas recognition non-linear model.

[0021] In any of the above possible implementation manners, a further implementation manner is provided, wherein the gas-sensitive material is SnO2, or SnO2 doped with NiO.

[0022] In any of the above possible implementation manners, a further implementation manner is provided, wherein the gas sensor element and the heating element are integrally prepared, and the specific preparation method is as follows:

[0023] On one side of the surface of the heating element, platinum paste is formed by screen printing and fired at 800 ± 10 °C to form interdigital electrodes; similarly, the gas-sensitive material paste is formed and sintered at 600 ± 10 °C for 4 h to prepare the thin-film type gas-sensitive element; the gas-sensitive material paste is a mixed paste of a gas-sensitive material, terpineol, and ethyl cellulose.

[0024] For any of the possible implementation manners described above, a further implementation manner is provided. Specifically, the gas recognition non-linear model is:

[0025] Multiple different temperature levels, temperature change processes, and atmospheres are set, and through experiments, the resistance values at different temperature levels and the temperature change following times under different temperature change processes are obtained under atmospheres of different concentrations to form training sample data; using the BP neural network algorithm and training the parameters with error backpropagation, the optimal BP neural network algorithm parameters are obtained.

[0026] For any of the possible implementation manners described above, a further implementation manner is provided. The sensor further includes a memory and a display. The microprocessor stores the measurement time, the temperature of the heating element, the resistance of the gas-sensitive element, and the atmosphere concentration data in the memory, or simultaneously displays the data on the display in real time.

[0027] For any of the possible implementation manners described above, a further implementation manner is provided. The temperature signal of the heating element is collected by Rt to obtain U x , and the working temperature of the heating element is calculated; the resistance signal of the gas-sensitive element is collected by SensingCell to obtain U L , and the resistance value of the gas-sensitive element is obtained; when controlling the temperature of the heating element, the supply voltage U(t) and the tracking voltage U H are respectively provided by both ends of a parallel high-value resistor.

[0028] For any of the possible implementation manners described above, a further implementation manner is provided. The temperature control of the heating element is performed by an adjustable buck chip, a digital rheostat, and a microprocessor to execute PID temperature control;

[0029] The PID temperature control adopts a discretized incremental form:

[0030] Δu n =K P (e n -e n-1 )+K I e n +K D (e n -2e n-1 +e n-2 );

[0031] Where e nis the temperature difference between the set temperature and the temperature collected at the nth time (T set - T n ), K P , K I and K D The parameter settings respectively consider the temperature control accuracy, the temperature overshoot, the adjustment time and the anti-interference ability, and calculate the voltage increment Δu n at the nth time.

[0032] The anti-interference comes from the working principle of PID, where P is proportional control, I is integral control, and D is derivative control. Simple proportional control is difficult to avoid steady-state error. Since the integral term of the error depends on the integral of time, the integral term will increase as time increases. By introducing integral control (I), the system can have almost no steady-state error after entering the steady state; in derivative control, the output of the controller is proportional to the derivative of the input error signal (i.e., the rate of change of the error). What is increased is the "derivative term", which can predict the trend of error change, so that the control action to suppress the error can be made equal to zero or even negative in advance, thus avoiding serious overshoot of the controlled quantity. For a controlled object with large inertia or lag, it can improve the dynamic characteristics of the system during the adjustment process.

[0033] In any of the possible implementation manners described above, a further implementation manner is provided. The gas recognition non-linear model adopts a BP neural network structure model, and based on the temperature level and the adsorption-desorption difference of the gas on the gas-sensitive element during the temperature change process, a non-linear relationship between the input data x i and the output data y q is established:

[0034]

[0035]

[0036] In the formula, w ij is the connection weight from the input layer to the hidden layer; θ j is the threshold of the hidden layer node; x i is the input data; v jq is the connection weight from the hidden layer to the output layer; h j is the output of the hidden layer node; γ q is the threshold of the output layer; p is the number of hidden layer nodes; y q is the output data; f(·) is the activation function of the BP neural network.

[0037] According to formulas (3) and (4), the gas type is identified and the corresponding gas concentration is calculated.

[0038] On the other hand, the present invention also provides a method for quantifying H2 and CO gases. The method uses the tin dioxide type sensor for quantifying H2 and CO gases described above, and the method includes:

[0039] S1. Establish a gas recognition non-linear model: Set multiple different temperature levels, temperature change processes, and atmospheres. Through experiments, obtain the resistance values at different temperature levels and the temperature change following time under different temperature change processes in atmospheres with different concentrations to form training sample data; Use the BP neural network algorithm to train parameters with error backpropagation to obtain the optimal BP neural network algorithm parameters;

[0040] S2. Place the sensor in the atmosphere to be measured. Convert the temperature signal on the heating element and the resistance signal on the gas-sensitive element into voltage signals U x and U L through a circuit, and convert them into signals recognizable by the microprocessor through an analog-to-digital converter; After being processed by the microprocessor, output the heating voltage u(t) to the control end of the heating element to meet the working temperature of the gas-sensitive element;

[0041] S3. According to the temperature signal on the heating element and the resistance signal on the gas-sensitive element obtained in step S2, use the gas recognition non-linear model obtained in step S1 to identify the type of gas to be measured and solve the corresponding concentration value.

[0042] For any of the possible implementation manners described above, a further implementation manner is provided. The method further includes:

[0043] S4. Store the measurement time, heating element temperature, gas-sensitive element resistance, gas type, and other concentration data in a memory, or simultaneously display the data on a display in real time.

[0044] The beneficial effects of the present invention are as follows: By integrating the thin-film gas-sensitive element and the heating element into one, and combining the incremental PID method for temperature control, the problem of temperature following of the gas-sensitive element during temperature rise and fall is solved, and at the same time, the anti-interference ability of the environment is improved; Utilize the differences in adsorption-desorption characteristics of gas temperature levels and their temperature changes, and establish a non-linear quantization relationship between the response value and the gas concentration with the help of the BP artificial neural network, thus solving the problem of gas recognition and quantification by the gas-sensitive material. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 Shown is a schematic diagram of the integration of the gas-sensitive element and the heating element in the embodiment.

[0046] Figure 2 Shown is a schematic diagram of signal acquisition, conversion, and its circuit in the embodiment.

[0047] Figure 3 Shown is a schematic diagram of the principle of incremental PID temperature control in the embodiment.

[0048] Figure 4 Shown is the BP neural network structure model diagram in the embodiment.

[0049] Figure 5 Shown is Ni x Sn 1-x O 2-x XRD spectrum of the gas-sensitive material.

[0050] Figure 6 Shown is the PID temperature control execution software program block diagram in the embodiment.

[0051] Figure 7 Shown is the influence diagram of the K value on the temperature and current on the heating element (a) and the gas-sensitive element (b) at 250 °C in the embodiment. P

[0052] Figure 8 Shown is the variation diagram of the resistance of SnO2 material with temperature under different CO and H2 atmospheres in the embodiment. Specific embodiments

[0053] The specific embodiments of the present invention will be described in detail below with reference to the specific drawings. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered as isolated, and they can be combined with each other to achieve better technical effects. In the drawings of the following embodiments, the same reference numerals appearing in each drawing represent the same features or components, and can be applied to different embodiments.

[0054] In an embodiment of the present invention, a tin dioxide type sensor for quantitatively detecting H2 and CO gases, the gas-sensitive element and the heating element are integrated into one body. The heating element is a flat plate structure (MCH, 10×10×1.2 mm) with a cermet embedded and an alumina outer cover, and the positive and negative electrodes are led out as the heating control terminals; the gas-sensitive element is a thin film formed by screen-printing SnO2 nano-powder and then firing at high temperature on one side of the alumina flat plate to enhance the followability of the resistance value of the gas-sensitive element to the temperature of the heating element. An inert metal interdigital electrode is sandwiched between them, and the positive and negative measurement terminals of the resistance of the gas-sensitive element are led out; a thermistor (Pt100) is arranged on the other side of the alumina flat plate to feedback the temperature signal of the gas-sensitive element, as Figure 1 shown.

[0055] In a specific embodiment, for the signal acquisition, conversion and display of the tin dioxide type sensor for quantitatively detecting H2 and CO gases, the temperature signal on the heating element and the resistance signal on the gas-sensitive element are converted into voltage signals (U x and U L ​) is converted into a signal recognizable by the microprocessor through an analog-to-digital converter. The signal is processed by the operation-control software, and the heating voltage u(t) is output to the control terminal of the heating element to meet the working temperature of the gas sensor element, such as Figure 2 shown.

[0056] In a specific embodiment, the tin dioxide type sensor for quantitatively detecting H2 and CO gases adjusts the proportional parameter (K P ), integral parameter (K I ), and differential parameter (K D ) in the PID temperature control model to meet the temperature control accuracy, temperature change rate, and anti-interference ability; based on the temperature difference of gas adsorption-desorption on the gas sensor element, a non-linear model for gas recognition is established to identify the gas type and solve the concentration value. The PID temperature control model adopts a discretized incremental form, that is, Δu n = K P (e n - e n-1 ) + K I e n + K D (e n - 2e n-1 + e n-2 ), where e n is the difference between the set temperature and the nth collected temperature (T set - T n ), and the parameters K P , K I , and K D are set mainly considering the temperature control accuracy, temperature overshoot, adjustment time, and anti-interference ability, and the nth output voltage increment Δu n is calculated, as shown in Figure 3 . For gas recognition and quantification, a BP neural network structure model is adopted, as shown in Figure 4 . Based on the temperature level and the difference in gas adsorption-desorption on the gas sensor element during the temperature change process, a non-linear relationship between the input data x i and the output data y q is established:

[0057]

[0058]

[0059] In the formula, w ij is the connection weight from the input layer to the hidden layer; θ j is the threshold of the hidden layer node; x i is the input data; v jq is the connection weight from the hidden layer to the output layer; h j is the output of the hidden layer node; γ qis the threshold of the output layer; p is the number of hidden layer nodes; y is the output data; f(·) is the activation function of the BP neural network. Thus, the gas type is identified and the corresponding gas concentration is calculated.

[0060] In a specific embodiment, for the tin dioxide type sensor for quantitatively detecting H2 and CO gases, data storage and display are carried out. Data storage means storing the measurement time, the temperature of the heating element, the resistance of the gas-sensitive element, the gas concentration, etc. in a USB flash drive, and at the same time displaying the instantaneous data on the liquid crystal screen. As Figure 2 shown, the resistance and temperature signals are converted into voltage signals, which are obtained by conversion through an analog-to-digital converter and the digital quantity is transmitted to the microprocessor for reading, and the corresponding relationship between the digital (N) and the voltage (U) is established, that is, U = f(N) or N = f(U) for conversion processing. For data storage and display, the measurement time, the temperature of the heating element, the resistance of the gas-sensitive element, and the quantitative calculation results of the gas components, etc. are stored in the USB flash drive through a serial interface or the instantaneous data is displayed on the liquid crystal module.

[0061] Embodiment 1

[0062] I. Preparation of gas-sensitive material

[0063] To obtain nano-scale powder, the sol-gel method is adopted to prepare a gel with a three-dimensional polymer or particle spatial structure of SnO2 and NiO-doped SnO2, and it is calcined at 500 °C for 2 h with a heating rate of 2 °C / min to obtain the gas-sensitive material, and its XRD spectrum is as Figure 5 shown.

[0064] II. Integration of heating element and gas-sensitive element

[0065] Using an alumina cermet sheet as the heating element (10×10×1.2 mm), platinum paste is formed on one side of the surface by screen printing and fired at 800 °C to form interdigital electrodes. Similarly, the gas-sensitive material slurry (gas-sensitive material + terpineol + ethyl cellulose) is formed and sintered at 600 °C for 4 h to form a thin-film gas-sensitive element, as Figure 1 shown in (a); on the other side of the surface, a thermistor (2×2×0.5 mm) is adhered as the temperature-measuring element, as Figure 1 shown in (b).

[0066] III. Signal acquisition and output

[0067] The temperature signal of the heating element (heating cell, Figure 2 ) is collected by Rt to obtain U x , and the working temperature is calculated. When Rt takes Pt100, its resistance

[0068] R T = 100×(1 + 3.908×10 -3 T - 5.802×10-7 T 2 ) (5)

[0069] Where T is the operating temperature, °C. The resistance signal of the gas sensor (sensing cell) is collected by SensingCell L , obtain its resistance value, and calculate the measured gas response value R=R a / R g , where R a , R g Corresponding to the resistance values of air and measured gas respectively. Heating element temperature control, supply voltage U(t) and tracking voltage U H , respectively provided by running the software-temperature control hardware and by connecting high-value resistors in parallel across the terminals.

[0070] 4. Temperature control of heating elements

[0071] The PID software temperature control program is executed by an adjustable step-down chip, a digital resistor and a microprocessor, such as Figure 6 As shown, the temperature control process is completed. Among them, the PID operation includes K P , K I and K D Parameter adjustment, such as Figure 7 As shown in (a), the heating element is heated from room temperature to 250°C to a constant temperature curve, and the time to reach a stable temperature is 17s. The corresponding current change on the SnO2 gas sensor in the air is as follows: Figure 7 As shown in (b), the change trends of current and resistance are consistent.

[0072] 5. Establishing the quantitative relationship between nonlinear temperature modulation parameters and concentration

[0073] Studies have shown that there is a nonlinear relationship between the adsorption response value of multi-component gases and their concentrations, and it is very difficult to establish an analytical expression. The present invention uses the BP (Back Propagation) neural network algorithm proposed by Rumelhart et al. in 1986, such as Figure 4 As shown, the parameters are trained by error back propagation, taking advantage of its good fault tolerance, adaptability and nonlinear processing characteristics.

[0074] Take five temperature levels, eight temperature change processes and 15 atmospheres, as shown in Table 1, where 0 is air and the airflow rate is 1.57×10 -2 m / s, typical experimental results are as follows Figure 8 From this, the resistance value of the constant temperature can be extracted (5: R 200 =R 1x ,R 250 =R 2x ,R 300 =R 3x,R 350 = R 4x ,R 400 = R 5x ) and the temperature change following time, i.e., the time difference between the change in resistance value and the change in temperature reaching 90% of the stable value during the heating and cooling process (8: t 200℃→250℃ = t 1x ,t 250℃→300℃ = t 2x ,t 300℃→350℃ = t 3x ,t 350℃→400℃ = t 4x ,t 400℃→350℃ = t 5x ,t 350℃→300℃ = t 6x ,t 300℃→250℃ = t 7x ,t 250℃→200℃ = t 8x ) to form matrix A. The subscript x = a, b,..., o corresponds to 15 groups of concentration matrices B, including H2, CO, i.e., the combined concentration, in ppm. Figure 8 Corresponding to the concentrations x = a, e, h, j, l, n.

[0075] Table 1 Experimental conditions for establishing the parameter matrix

[0076]

[0077] Matrices A and B are the training sample data (i.e., the input data), obtained from the above experimental data. According to the BP neural network structure model ( Figure 4 ), it is transformed through the activation function. In equations (3) and (4), the activation function f() takes Tanh:

[0078]

[0079] It is optimized through the error function:

[0080]

[0081] Train the weights w ij and v jq , until the set error requirement is met.

[0082]

[0083]

[0084] For the SnO2 material, set the relative error to 2%, take 10 hidden layer neurons, and obtain w ij and v jq Coefficient matrix:

[0085]

[0086] and

[0087]

[0088] and

[0089]

[0090]

[0091] Then, through the resistance value and the following time matrix C under the temperature-varying mode, the matrix D (output data) of the gas concentration to be determined is obtained.

[0092]

[0093]

[0094] To verify its feasibility, it is respectively applied to SnO2 and NiO-doped SnO2 materials. As shown in Table 2, in the range of 0 - 1000 ppm of H2, CO, and H2+CO mixed gases, the maximum relative error is only 2.32%.

[0095] Table 2 Prediction results using different gas-sensitive materials

[0096]

[0097] The gas-sensitive element, heating element, and thermistor of the present invention are integrated, satisfying the rapid response of the response signal to temperature changes; the combination of PID temperature control and signal acquisition module obtains a fast and stable temperature environment, and at the same time has strong anti-interference ability; in particular, it can identify gases by using the differences in gas adsorption, solving the selectivity of such sensors. Finally, through the combination of signal acquisition, conversion, and liquid crystal screen, the detection results are given quickly and intuitively.

[0098] Although several embodiments of the present invention have been given in this article, those skilled in the art should understand that the embodiments in this article can be changed without departing from the spirit of the present invention. The above embodiments are only exemplary and should not be used as the limitation of the scope of the rights of the present invention.

Claims

1. A tin dioxide type sensor for quantitatively detecting H2 and CO gases, characterized in that, The sensor includes a heating element, a gas-sensitive element, a thermosensitive element, and a microprocessor; The heating element has a flat structure with a cermet-inside and alumina-outside housing, and the positive and negative electrodes are led out as the heating control terminals; The gas-sensitive element is a thin film formed by screen-printing a nano-powder of a gas-sensitive material and then firing it at a high temperature on one side of the heating element; interdigital electrodes are arranged between the gas-sensitive element and the heating element, and the positive and negative measurement terminals of the gas-sensitive element are led out; The thermosensitive element is arranged on the other side of the heating element to feedback the temperature signal of the gas-sensitive element; The microprocessor controls the temperature change of the heating element, receives the signals of the gas-sensitive element and the thermosensitive element, and calculates the type and concentration value of the detected gas according to the gas recognition non-linear model; The gas-sensitive element and the heating element are integrally prepared, and the specific preparation method is as follows: On one side of the surface of the heating element, platinum paste is formed by screen-printing and fired at 800±10°C to form interdigital electrodes; similarly, the gas-sensitive material paste is formed and sintered at 600±10°C for 4h to prepare the thin-film type gas-sensitive element; the gas-sensitive material paste is a mixed paste of a gas-sensitive material, terpineol, and ethyl cellulose; The gas recognition non-linear model is specifically as follows: Set multiple different temperature levels, temperature change processes, and atmospheres, and obtain the resistance values at different temperature levels and the temperature change following time under different temperature change processes in atmospheres with different concentrations through experiments to form training sample data; use the BP neural network algorithm to train the parameters with error backpropagation to obtain the optimal BP neural network algorithm parameters; the obtained optimal BP neural network model is the gas recognition non-linear model.

2. The tin dioxide type sensor for quantitatively detecting H2 and CO gases according to claim 1, characterized in that, The gas-sensitive material is SnO2, or SnO2 doped with NiO.

3. The tin dioxide type sensor for quantitatively detecting H2 and CO gases according to claim 1, characterized in that, The sensor further includes a memory and a display. The microprocessor stores the measurement time, heating element temperature, gas-sensitive element resistance, and atmosphere concentration data in the memory, or simultaneously displays the data on the display in real time.

4. The tin dioxide type sensor for quantitatively detecting H2 and CO gases according to claim 1, characterized in that, The temperature signal of the heating element is used to collect the voltage U of the heating element through Rt x , and the working temperature of the heating element is calculated; the resistance signal of the gas sensor element is used to collect the voltage U of the gas sensor element through SensingCell L , and the resistance value of the gas sensor element is obtained; when controlling the temperature of the heating element, the supply voltage U(t) and the tracking voltage U H are respectively provided through both ends of a parallel high-value resistor.

5. The tin dioxide type sensor for quantitatively detecting H2 and CO gases according to claim 4, characterized in that, The temperature control of the heating element is performed by an adjustable buck chip, a digital varistor, and a microprocessor to execute PID temperature control; The PID temperature control adopts a discretized incremental form: Δu n = K P (e n - e n-1 ) + K I e n + K D (e n - 2e n-1 + e n-2 ); where e n is the temperature difference between the set temperature and the temperature collected at the nth time (T set -T n ), and K P , K I and K D The parameter settings respectively consider the temperature control accuracy, the temperature overshoot, the adjustment time and the anti-interference ability, and calculate the output voltage increment Δu n .

6. The tin dioxide type sensor for quantitatively detecting H2 and CO gases according to claim 1, characterized in that, The gas recognition non-linear model adopts a BP neural network structure model, and based on the temperature level and the adsorption-desorption difference of the gas on the gas-sensitive element during the temperature change process, the input data x i and the output data y q The non-linear relationship between the two is: where, w ij is the connection weight from the input layer to the hidden layer; θ j is the threshold of the hidden layer node; x i is the input data; v jq is the connection weight from the hidden layer to the output layer; h j is the output of the hidden layer node; γ q is the threshold of the output layer; p is the number of hidden layer nodes; m is the total number of input data; y q is the output data; f(·) is the activation function of the BP neural network; Identify the gas type and calculate the corresponding gas concentration according to formulas (3) and (4).

7. A method for quantifying H2 and CO gases, characterized in that, The method uses a tin dioxide type sensor for quantitatively detecting H2 and CO gases as described in any one of claims 1-6. The method includes: S1. Establish a gas recognition non-linear model: Set multiple different temperature levels, temperature change processes, and atmospheres, and obtain the resistance values at different temperature levels and the temperature change following time under different temperature change processes in atmospheres with different concentrations through experiments to form training sample data; use the BP neural network algorithm to train the parameters with error backpropagation to obtain the optimal BP neural network algorithm parameters; the obtained optimal BP neural network model is the gas recognition non-linear model; S2. Place the sensor in the atmosphere to be measured. Convert the temperature signal on the heating element and the resistance signal on the gas sensor element into voltage signals respectively through a circuit, and then convert them into signals recognizable by the microprocessor through an analog-to-digital converter. After being processed by the microprocessor, output the heating voltage u(t) to the control terminal of the heating element to meet the operating temperature of the gas sensor element. S3. According to the temperature signal on the heating element and the resistance signal on the gas sensor element obtained in step S2, use the gas recognition non-linear model obtained in step S1 to identify the type of gas to be measured and solve the corresponding concentration value.

8. The method for quantifying H2 and CO gases according to claim 7, characterized in that, The method further includes: S4. Store the measurement time, heating element temperature, gas sensor element resistance, gas type, and other concentration data in the memory, or simultaneously display the data on the display in real time.

Images