Gas thermophysical parameter identification method and system based on MEMS sensor

Abstract

The present application relates to the technical field of gas detection and thermal physical analysis, in particular to a gas thermal physical parameter identification method and system based on a MEMS sensor. The method comprises: applying a multi-frequency alternating excitation current to a heating element to make it generate a two-fold reference frequency Joule heat fluctuation and cause a resistance same-frequency fluctuation; extracting a three-fold frequency output voltage signal through a lock-in amplifier circuit, decomposing it into in-phase and quadrature components by using phase-sensitive detection, then calculating a complex temperature rise and constructing a complex thermal impedance model; calculating the thermal conductivity based on the slope of the real part signal of the complex thermal impedance model changing with frequency, extracting the thermal diffusivity from the imaginary part and phase delay; and finally decoupling and outputting the volume specific heat capacity in combination with a correlation model. The present application realizes high-precision decoupling of multi-dimensional thermal parameters through three-fold frequency domain impedance analysis, effectively filters out drift interference, and solves the problem of difficult accurate identification of complex component gases at micro scale.

Description

Technical Field

[0001] This invention relates to the field of gas detection and thermal property analysis technology, specifically to a method and system for identifying gas thermal property parameters based on MEMS sensors. Background Technology

[0002] With the development of the Industrial Internet of Things and the hydrogen economy, the accurate identification of thermophysical parameters of mixed gases in complex environments has become a key technological requirement for energy security monitoring and quantitative analysis of gas components.

[0003] Traditional gas detection primarily relies on chemical sensing techniques, but these are prone to drift and poisoning under high temperature and humidity conditions. Currently, MEMS-based thermal conductivity sensors achieve physical detection by measuring the intrinsic thermal parameters of gases, exhibiting better long-term stability. However, most existing identification methods employ steady-state DC drive or simple transient pulse methods, only obtaining a single thermal conductivity coefficient, making it difficult to effectively separate multiple physical parameters in complex background gases. When dealing with gas components with similar thermal conductivity (such as carbon dioxide and argon) or multi-component mixtures (such as hydrogen-mixed natural gas), existing technologies lack the ability to decouple multi-dimensional physical quantities, resulting in insufficient identification accuracy and severe cross-sensitivity, failing to meet the requirements of high-precision qualitative and quantitative analysis in industrial applications.

[0004] The technical problem to be solved by this invention is: how to improve the accuracy and decoupling capability of simultaneous identification of multiple intrinsic thermal parameters such as thermal conductivity, thermal diffusivity and volumetric specific heat capacity in a multi-component mixed gas environment.

[0005] To address this, a method and system for identifying gas thermal properties based on MEMS sensors are proposed. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for identifying gas thermophysical parameters based on MEMS sensors. By using third harmonic frequency domain impedance analysis, high-precision decoupling of multidimensional thermal parameters is achieved, solving the problem of accurate identification of complex gas components.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] Methods for identifying gas thermal properties based on MEMS sensors include:

[0009] A multi-frequency AC excitation current is applied to the heating element of the MEMS sensor, causing the heating element to generate Joule thermal fluctuations at twice the reference frequency, resulting in the resistance value of the heating element fluctuating at the same frequency.

[0010] The third harmonic output voltage signal generated across the heating element is detected and extracted in real time by a locked amplifier circuit; the third harmonic output voltage signal is decomposed into an in-phase component signal that is in phase with the AC excitation current and an orthogonal component signal that has a phase deviation from the AC excitation current using phase-sensitive detection technology; the complex temperature rise of the heating element is calculated based on the in-phase component signal and the orthogonal component signal, and a complex thermal impedance model of the gas under test is constructed by combining the Joule thermal fluctuation;

[0011] The thermal conductivity of the gas under test is calculated based on the slope of the real part of the complex thermal impedance model as a function of the reference frequency; the thermal diffusivity of the gas under test is extracted based on the imaginary part of the complex thermal impedance model and the phase delay.

[0012] By utilizing the thermal conductivity and thermal diffusivity, combined with the correlation model, the volumetric specific heat capacity of the gas under test is decoupled and output, and the component ratio of the mixed gas is quantitatively analyzed.

[0013] Preferably, the step of applying multi-frequency AC excitation current includes: generating n sets of reference sine wave signals with different preset frequencies using a digital frequency synthesizer, and linearly weighting and superimposing the n sets of reference sine wave signals to synthesize a composite signal containing n target reference frequency components; inputting the composite signal to a constant current drive circuit, and outputting the multi-frequency AC excitation current to the heating element through closed-loop control; adjusting the weight of each reference frequency in the multi-frequency AC excitation current according to the thermal response feedback of the gas to be tested, so that the resistance value generates same-frequency fluctuations containing n components twice the reference frequency, and generates a triple-frequency output voltage signal corresponding to each of the reference frequencies.

[0014] Preferably, the specific steps for generating the third-harmonic output voltage signal include: the multi-frequency AC excitation current causing the heating element to generate temperature fluctuations containing twice the components of each reference frequency, and linearly modulating the temperature fluctuations into a dynamic resistance signal containing twice the components of the reference frequency based on the temperature coefficient of resistance; according to Ohm's law, the multi-frequency AC excitation current with a frequency of reference frequency interacts with the dynamic resistance signal with a frequency of twice the reference frequency across the heating element to generate an original voltage signal containing a component of three times the reference frequency; the original voltage signal is processed by the lock-in amplifier circuit to filter out the reference frequency components and extract the third-harmonic output voltage signal characterizing the thermophysical properties of the gas under test.

[0015] Preferably, the specific steps for extracting and decomposing the in-phase component signal and the quadrature component signal include: using a digital frequency synthesizer to generate a first reference sine signal that is in phase and has the same frequency as three times the reference frequency, and a second reference sine signal that has a 90° phase difference with the first reference sine signal; using phase-sensitive detection technology, multiplying the third harmonic output voltage signal with the first reference sine signal and the second reference sine signal respectively, modulating the target frequency signal to the zero frequency band, and generating a mixed signal containing twice the three times the reference frequency; using a low-pass filter to filter the mixed signal, filtering out high-frequency components, and extracting in real time the in-phase component signal representing the real part of the third harmonic output voltage signal, and the quadrature component signal representing the imaginary part of the output voltage signal.

[0016] Preferably, the steps for constructing a complex thermal impedance model of the gas under test include: using the in-phase component signal, the quadrature component signal, and the resistance temperature coefficient of the heating element, calculating the complex temperature rise amplitude of the heating element at twice the reference frequency and its phase lag relative to the AC excitation current; calculating the Joule thermal fluctuation power generated by the heating element based on the effective value of the AC excitation current and the resistance base value of the heating element, defining it as the dynamic heat flux vector input to the gas environment under test; and performing a complex division operation between the complex temperature rise amplitude and the dynamic heat flux vector to construct the complex thermal impedance model.

[0017] Preferably, the steps for calculating thermal conductivity and thermal diffusivity include: extracting the real part signal of the complex thermal impedance model within a preset frequency range, establishing a linear mapping relationship between the real part signal and the reference frequency; calculating the slope of the linear mapping relationship, and inverting the thermal conductivity of the gas to be tested by combining the geometric characteristic parameters of the heating element; and using the imaginary part signal of the complex thermal impedance model, combined with the thermal conductivity, calculating and extracting the thermal diffusivity of the gas to be tested by intercept fitting.

[0018] Preferably, the step of outputting the volumetric specific heat capacity includes: establishing a thermal parameter characteristic matrix of the gas to be tested based on the physical correlation formula between thermal diffusivity, thermal conductivity, and volumetric specific heat capacity; the physical correlation formula satisfies the mathematical relationship that thermal diffusivity equals thermal conductivity divided by volumetric specific heat capacity; substituting the calculated thermal conductivity and the extracted thermal diffusivity into the physical correlation formula, and decoupling the volumetric specific heat capacity of the gas to be tested through numerical division.

[0019] A gas thermophysical property parameter identification system based on MEMS sensors includes:

[0020] Drive module: applies multi-frequency AC excitation current to the heating element of the MEMS sensor; causes the heating element to generate Joule thermal fluctuations at twice the reference frequency, resulting in the resistance value of the heating element fluctuating at the same frequency;

[0021] Extraction module: Real-time detection and extraction of the third harmonic output voltage signal generated across the heating element via a locked amplification circuit; decomposition of the third harmonic output voltage signal into an in-phase component signal that is in phase with the AC excitation current and an orthogonal component signal that has a phase deviation from the AC excitation current using phase-sensitive detection technology; calculation of the complex temperature rise of the heating element based on the in-phase component signal and the orthogonal component signal; and construction of a complex thermal impedance model of the gas under test based on the Joule thermal fluctuations.

[0022] Calculation module: Calculates the thermal conductivity of the gas under test based on the slope of the real part signal of the complex thermal impedance model as a function of the logarithm of the reference frequency; extracts the thermal diffusivity of the gas under test based on the imaginary part signal and phase delay of the complex thermal impedance model.

[0023] Analysis module: Utilizing the thermal conductivity and thermal diffusivity, combined with the correlation model, the volumetric specific heat capacity of the gas under test is decoupled and output to quantitatively analyze the component ratio of the mixed gas.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0025] 1. This invention utilizes third-harmonic frequency domain impedance analysis technology, leveraging the nonlinear electrothermal response of a heating element under multi-frequency AC excitation, to simultaneously extract multiple intrinsic physical parameters such as thermal conductivity, thermal diffusivity, and volumetric specific heat capacity in a single experiment. This multi-parameter collaborative identification mechanism provides auxiliary criteria for distinguishing gases with similar thermal properties, significantly improving the selectivity and classification accuracy of the sensor in complex background gases. By establishing a deep mapping between macroscopic thermal flow fields and microscopic molecular characteristics, this invention achieves a leap from single-physical quantity sensing to full-dimensional thermal property identification, greatly enhancing the reliability of qualitative and quantitative analysis of multi-component mixed gases.

[0026] 2. This invention employs a composite excitation signal composed of multiple superimposed reference frequencies, combined with closed-loop constant current drive and dynamic weight adjustment technology, effectively solving the problem of weak and easily interfered electrothermal signals at the microscale. By extracting the in-phase and quadrature components of the third harmonic output voltage through phase-sensitive detection technology, the system can effectively filter out DC drift, power frequency interference, and background electromagnetic noise, resulting in an order-of-magnitude improvement in detection limit and recognition accuracy. Simultaneously, the constructed complex thermal impedance model can cancel common-mode interference from ambient temperature and pressure, significantly enhancing zero-point stability over a wide temperature range and in complex dynamic scenarios.

[0027] 3. This invention utilizes the intrinsic physical correlation formula between thermal conductivity, thermal diffusivity, and volumetric specific heat capacity to directly decouple the energy characteristic parameters of a gas through numerical calculations. Compared to pure machine learning models, this physical mechanism-based decoupling algorithm offers higher transparency and universality, enabling rapid adaptation to fluctuations in energy composition across different production sites. Furthermore, this method requires extremely low system temperature rise, and combined with an event-driven sampling mechanism, it provides a core technological foundation for achieving edge intelligent monitoring and ultra-long-life wireless sensor networks. Attached Figure Description

[0028] Figure 1 This is a flowchart of the gas thermophysical parameter identification method based on MEMS sensors proposed in this invention.

[0029] Figure 2 This is a flowchart illustrating the gas thermophysical parameter identification method based on MEMS sensors proposed in this invention.

[0030] Figure 3 This is a system structure diagram of the gas thermophysical parameter identification system based on MEMS sensors proposed in this invention. Detailed Implementation

[0031] 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.

[0032] Example 1

[0033] Please see Figures 1 to 2 This invention provides a method for identifying gas thermal property parameters based on MEMS sensors, and the technical solution is as follows:

[0034] Methods for identifying gas thermal properties based on MEMS sensors, such as Figures 1-2 As shown, it includes:

[0035] A multi-frequency AC excitation current is applied to the heating element of the MEMS sensor, causing the heating element to generate Joule thermal fluctuations at twice the reference frequency, resulting in the resistance value of the heating element fluctuating at the same frequency.

[0036] The third harmonic output voltage signal generated across the heating element is detected and extracted in real time by a locked amplifier circuit; the third harmonic output voltage signal is decomposed into an in-phase component signal that is in phase with the AC excitation current and an orthogonal component signal that has a phase deviation from the AC excitation current using phase-sensitive detection technology; the complex temperature rise of the heating element is calculated based on the in-phase component signal and the orthogonal component signal, and a complex thermal impedance model of the gas under test is constructed by combining the Joule thermal fluctuation;

[0037] The thermal conductivity of the gas under test is calculated based on the slope of the real part of the complex thermal impedance model as a function of the reference frequency; the thermal diffusivity of the gas under test is extracted based on the imaginary part of the complex thermal impedance model and the phase delay.

[0038] By utilizing the thermal conductivity and thermal diffusivity, combined with the correlation model, the volumetric specific heat capacity of the gas under test is decoupled and output, and the component ratio of the mixed gas is quantitatively analyzed.

[0039] Further, the step of applying multi-frequency AC excitation current includes: generating n sets of reference sine wave signals with different preset frequencies using a digital frequency synthesizer, and linearly weighting and superimposing the n sets of reference sine wave signals to synthesize a composite signal containing n target reference frequency components; inputting the composite signal to a constant current drive circuit, and outputting the multi-frequency AC excitation current to the heating element through closed-loop control; adjusting the weight of each reference frequency in the multi-frequency AC excitation current according to the thermal response feedback of the gas to be tested, so that the resistance value generates same-frequency fluctuations containing n components twice the reference frequency, and generates a triple-frequency output voltage signal corresponding to each of the reference frequencies.

[0040] The steps for adjusting the weights based on the thermal response feedback of the gas under test include: real-time monitoring of the amplitude of each third harmonic output voltage signal component and calculating the signal-to-noise ratio (SNR) at each frequency point; if the SNR of the low-frequency signal is lower than a preset threshold due to environmental vibration and noise interference, the gain weight of the corresponding low-frequency reference sine wave signal in the composite signal is increased; if the high-frequency signal is affected by parasitic capacitance feedthrough effect, the weight of the high-frequency reference signal is reduced and the phase compensation parameters of the constant current drive circuit are adjusted to ensure that the complex temperature rise generated by each frequency component remains amplitude-normalized within the dynamic range of the lock-in amplifier. In actual identification, the preset threshold for the SNR is typically set to 20dB to 40dB.

[0041] The frequency distribution of the n sets of reference sinusoidal signals follows the principle of non-harmonic correlation, avoiding the overlap of nonlinear offset components generated by different frequency components; the constant current drive circuit has high bandwidth linearity, ensuring that the intermodulation distortion between the frequency components of the composite signal is lower than the preset level during the electrothermal conversion process, thereby ensuring that each third harmonic output voltage signal can independently characterize the gas thermal property fingerprint at the corresponding frequency; the preset level of intermodulation distortion is usually set to be lower than -60dB (or less than 0.1%); the intermodulation distortion is characterized by the component crosstalk level of the third harmonic output voltage signal after phase-sensitive detection.

[0042] This embodiment improves identification reliability in complex industrial environments through multi-frequency weight dynamic adjustment and non-harmonic frequency layout. Signal-to-noise ratio feedback compensation for low-frequency vibration noise and high-frequency capacitive feedthrough effect ensures amplitude normalization and high fidelity across the entire frequency band. Non-harmonic correlation principles and high-bandwidth linear drive effectively suppress intermodulation distortion, eliminate crosstalk between frequencies, and guarantee independent extraction of each thermal property fingerprint.

[0043] Further, the specific steps for generating the third harmonic output voltage signal include: the multi-frequency AC excitation current causes the heating element to generate temperature fluctuations containing twice the components of each reference frequency, and the temperature fluctuations are linearly modulated into a dynamic resistance signal containing twice the reference frequency components based on the temperature coefficient of resistance; according to Ohm's law, the multi-frequency AC excitation current with a frequency of reference frequency and the dynamic resistance signal with a frequency of twice the reference frequency interact at both ends of the heating element to generate an original voltage signal containing a component of three times the reference frequency; the original voltage signal is processed by the lock-in amplifier circuit to filter out the reference frequency components and extract the third harmonic output voltage signal characterizing the thermophysical properties of the gas under test.

[0044] To ensure that the dynamic resistance signal can accurately track the temperature fluctuation, the highest reference frequency of the multi-frequency AC excitation current must be lower than the thermal relaxation frequency of the heating element and the surrounding gas diffusion layer; within the thermal penetration depth range, the Joule heat generated by the heating element is instantaneously transferred to the gas medium under test, establishing a quasi-steady-state thermal wave diffusion field.

[0045] The physical premise of the linear modulation lies in pre-controlling the amplitude of the AC excitation current of the heating element, so that the operating temperature rise of the heating element is maintained within the linear range of the material's resistance temperature characteristic, and the temperature rise amplitude is much smaller than the ambient absolute temperature, thereby eliminating the interference of higher-order nonlinear terms on the extraction of the third harmonic component. Furthermore, the phase-sensitive detection reference signal of the lock-in amplifier circuit maintains strict phase synchronization with the multi-frequency AC excitation current. By eliminating phase shifts caused by circuit feedthrough and parasitic impedance, it ensures that the extracted third harmonic output voltage signal accurately corresponds to the thermal conductivity and thermal diffusivity characteristics of the gas under test. The linear range refers to limiting the effective value of the multi-frequency AC excitation current so that the instantaneous temperature rise amplitude of the heating element during the measurement process is between 1K and 20K. Within this range, the resistance temperature coefficient fluctuation of the heating element is less than 0.1%.

[0046] This embodiment achieves high-precision extraction of weak thermophysical property signals by constructing a deep modulation mechanism involving electro-thermal-electricity. Utilizing the interaction between a second-harmonic dynamic resistor and the fundamental frequency current, the signal carrying gas characteristics is modulated to the third harmonic at the physical level, effectively avoiding interference from the fundamental frequency background signal and DC drift. By limiting the thermal relaxation frequency and linear temperature rise range, the sensor's response consistency during microscale heat transfer is ensured, enabling the output signal to accurately and linearly reflect the intrinsic parameters of the gas. Combined with rigorous phase synchronization technology, the sensitivity for identifying complex gas components is improved, providing physical support for achieving highly stable thermophysical property analysis with extremely low power consumption.

[0047] Further, the specific steps for extracting and decomposing the in-phase and quadrature component signals include: using a digital frequency synthesizer to generate a first reference sine signal that is in phase and has the same frequency as three times the reference frequency, and a second reference sine signal that has a 90° phase difference with the first reference sine signal; using phase-sensitive detection technology, multiplying the third harmonic output voltage signal with the first and second reference sine signals respectively, modulating the target frequency signal to the zero frequency band, and generating a mixed signal containing twice the three times the reference frequency; using a low-pass filter to filter the mixed signal, filtering out high-frequency components, and extracting in real time the in-phase component signal representing the real part of the third harmonic output voltage signal, and the quadrature component signal representing the imaginary part of the output voltage signal.

[0048] The phase-sensitive detection technology adopts a parallel quadrature demodulation architecture, using n independent digital multipliers to simultaneously perform calculations on the third harmonic output voltage signal and the third harmonic reference signal corresponding to each reference frequency. By setting the clock synchronization of the digital frequency synthesizer, it is ensured that each frequency component maintains strict frequency orthogonality during the quadrature demodulation process, thereby eliminating crosstalk between different frequency components.

[0049] The low-pass filter's parameter configuration includes: the cutoff frequency of the low-pass filter is adaptively set based on the minimum frequency difference between adjacent reference frequencies in the multi-frequency AC excitation current, and the cutoff frequency must be less than half of the minimum frequency difference to effectively suppress high-frequency step components and beat frequency interference generated by adjacent frequencies in the mixed signal; simultaneously, the order of the low-pass filter is optimized according to the system's dynamic response requirements to ensure that stable in-phase and quadrature component signals are output in real time while filtering out harmonic noise. Furthermore, combined with a magnitude normalization step, the gain compensation for in-phase and quadrature components at different frequencies is performed using a pre-calibrated amplitude-frequency characteristic curve of the circuit system, eliminating the amplitude nonlinear attenuation of the analog front-end circuit over a wide bandwidth.

[0050] This embodiment achieves efficient and clean extraction of multi-frequency thermal response signals through parallel orthogonal demodulation and adaptive filtering techniques. By utilizing frequency orthogonality to eliminate intermodulation interference between components, and in conjunction with a cutoff frequency set based on the frequency difference, high-frequency noise is thoroughly filtered out while precise thermal characteristic data is preserved. Furthermore, a circuit gain compensation mechanism is introduced to correct amplitude attenuation over a wide bandwidth, improving the calculation accuracy of in-phase and quadrature components.

[0051] Further, the steps for constructing the complex thermal impedance model of the gas under test include: using the in-phase component signal, the quadrature component signal, and the resistance temperature coefficient of the heating element, calculating the complex temperature rise amplitude of the heating element at twice the reference frequency and its phase lag relative to the AC excitation current; calculating the Joule thermal fluctuation power generated by the heating element based on the effective value of the AC excitation current and the resistance base value of the heating element, defining it as the dynamic heat flux vector input to the gas environment under test; and performing a complex division operation between the complex temperature rise amplitude and the dynamic heat flux vector to construct the complex thermal impedance model.

[0052] The process of constructing the complex thermal impedance model also includes a substrate parasitic heat flow subtraction step: the vacuum complex thermal impedance of the heating element is measured in advance in a vacuum environment to characterize the parasitic heat flow conducted through the sensor support structure and substrate; when measuring the gas to be measured, the calculated total complex thermal impedance is combined with the vacuum complex thermal impedance to perform a complex parallel cancellation operation, eliminating the substrate heat conduction component, and obtaining the pure gas to be measured complex thermal impedance.

[0053] Furthermore, the complex thermal impedance model undergoes geometric factor correction: a shape correction coefficient is introduced to calibrate the complex division result by incorporating the linewidth, line length, and effective thickness of the heat-affected zone of the heating element; in the calculation of phase lag, a compensation circuit is used to transmit the parasitic phase shift generated by the transmission line, ensuring that the phase lag is caused only by the thermal diffusion delay of the gas under test. Through these corrections, the complex thermal impedance model can accurately characterize the dynamic heat transfer resistance and thermal capacity response characteristics of the gas under test within a microscale chamber.

[0054] This embodiment significantly improves the absolute accuracy of gas thermal property identification through vacuum background cancellation and geometric factor correction. The use of a complex parallel model effectively eliminates parasitic heat losses from the sensor substrate and support structure, ensuring that the thermal impedance data is modulated only by the intrinsic characteristics of the gas being measured. Simultaneously, by compensating for parasitic phase shifts in the transmission lines and introducing a shape correction coefficient, deviations caused by sensor physical dimensions and circuit delays are eliminated. This improvement not only enhances the model's versatility across different hardware platforms but also ensures the reliability of characterizing gas thermal resistance and heat capacity features in a microscale environment.

[0055] Further, the steps for calculating the thermal conductivity and thermal diffusivity include: extracting the real part signal of the complex thermal impedance model within a preset frequency range, establishing a linear mapping relationship between the real part signal and the reference frequency; calculating the slope of the linear mapping relationship, and inverting the thermal conductivity of the gas to be tested by combining the geometric characteristic parameters of the heating element; and using the imaginary part signal of the complex thermal impedance model, combined with the thermal conductivity, calculating and extracting the thermal diffusivity of the gas to be tested by intercept fitting.

[0056] The selection of the preset frequency band is based on the matching relationship between the thermal wave penetration depth and the geometric dimensions of the space to be measured: the lower limit of the reference frequency is set to ensure that the thermal wave penetration depth is less than the characteristic dimensions of the sensor chamber, thereby eliminating boundary reflection interference; the upper limit of the reference frequency is set to meet the quasi-steady-state thermal conduction conditions, ensuring that the real part signal and the logarithm of the reference frequency are highly linear.

[0057] The specific inversion logic of thermal conductivity and thermal diffusivity includes: using the slope of the linear mapping relationship, combined with the heating power per unit length of the heating element, the thermal conductivity of the gas to be measured is obtained by inversion based on the analytical solution of the infinitely long linear heat source model; when extracting thermal diffusivity, the imaginary part signal of the complex thermal impedance model is used as a function characteristic of frequency variation, combined with the known thermal conductivity, and the intercept of the imaginary part response curve on the logarithmic axis of frequency is fitted, and substituted into the characteristic time constant formula containing thermal diffusivity, thereby eliminating the phase shift error caused by contact thermal resistance and realizing the accurate extraction of the thermal diffusivity of the gas to be measured.

[0058] This embodiment ensures the physical rigor and computational accuracy of thermal property parameter extraction through frequency boundary locking and analytical model inversion. Utilizing the thermal wave penetration depth to define the frequency range eliminates boundary interference and guarantees the reliability of the linear mapping of the real part signal. Introducing an infinitely long linear heat source model and a characteristic time constant achieves high-precision decoupling of thermal conductivity and thermal diffusivity, effectively eliminating the interference of contact thermal resistance on the phase signal.

[0059] Further, the step of outputting the volumetric specific heat capacity includes: establishing a thermal parameter characteristic matrix of the gas to be tested based on the physical correlation formula between thermal diffusivity, thermal conductivity, and volumetric specific heat capacity; the physical correlation formula satisfies the mathematical relationship that thermal diffusivity equals thermal conductivity divided by volumetric specific heat capacity; substituting the calculated thermal conductivity and the extracted thermal diffusivity into the physical correlation formula, and decoupling the volumetric specific heat capacity of the gas to be tested through numerical division.

[0060] Specifically, in order to suppress the accumulation of errors during the parameter decoupling process, the step of decoupling the output volumetric specific heat capacity further includes: using the weighted least squares method to verify the consistency of thermal conductivity and thermal diffusivity at multiple reference frequencies; establishing a parameter estimation model that includes the measurement noise covariance, and minimizing the residuals of the decoupled volumetric specific heat capacity, thermal conductivity, and thermal diffusivity under the physical correlation formula through iterative optimization, thereby eliminating the influence of fluctuations at a single frequency point on the accuracy of specific heat capacity calculation.

[0061] Furthermore, the specific steps for quantitatively analyzing the component ratios of the mixed gas include: constructing a three-dimensional thermal feature vector, whose components are the thermal conductivity, the thermal diffusivity, and the volumetric specific heat capacity; inputting the feature vector into a preset component mapping model, which is established based on the linear weighted thermophysical property combination law of multi-component gases; and calculating the offset of the gas component to be tested relative to the pure background gas by searching the position of the feature vector in the thermal parameter feature matrix space, thereby realizing the quantitative inversion of the concentration of each component in the mixed gas.

[0062] This embodiment improves the accuracy and reliability of mixed gas component analysis by using multi-source parameter fusion estimation and multi-dimensional feature mapping. The residual minimization algorithm effectively suppresses error propagation during the calculation process, ensuring that the decoupled volumetric specific heat capacity possesses extremely high numerical stability. By constructing three-dimensional feature vectors for thermal conductivity, thermal diffusivity, and specific heat capacity, the unique feature point of the mixed gas in the physical parameter space can be accurately located, eliminating the degeneracy and cross-interference inherent in single-parameter analysis.

[0063] Specifically, the correlation model includes a first physical correlation formula composed of thermal diffusivity, thermal conductivity, and volumetric specific heat capacity, and a second component mapping model for inverting component proportions. The first physical correlation formula strictly follows the law of conservation of energy transport, meaning the volumetric specific heat capacity of the gas being tested is equal to the ratio of thermal conductivity to thermal diffusivity. In the decoupled calculation, the system uses multiple sets of collected reference frequency response data and the first physical correlation formula to initially obtain the volumetric specific heat capacity value of the gas being tested, thereby constructing a three-dimensional intrinsic thermal characteristic vector including thermal conductivity, thermal diffusivity, and volumetric specific heat capacity.

[0064] The component mapping model is established based on a pre-calibrated thermal database of various pure gases and their known proportions of mixtures. Its input is the three-dimensional intrinsic thermal eigenvector, and its output is the volume fraction or mole fraction of the target component in the gas mixture to be tested. In the actual quantitative inversion process, the offset relative to the pure background gas is defined as the Euclidean distance of the three-dimensional eigenvector of the gas mixture to be tested from the coordinate points of the pure background gas in the feature space. The system searches for the coordinate position of the eigenvector in the thermal parameter feature matrix space, and uses a linear weighted mixing criterion or a preset lookup table algorithm to calculate the correspondence between the offset and the component concentration, thereby accurately inverting the percentage concentration of each component in the gas mixture.

[0065] This invention utilizes third-harmonic frequency domain impedance analysis technology and the nonlinear electrothermal response of a heating element under multi-frequency AC excitation to simultaneously extract multiple intrinsic thermophysical parameters, such as thermal conductivity, thermal diffusivity, and volumetric specific heat capacity, in a single measurement. This multi-parameter collaborative identification mechanism establishes a deep mapping between the macroscopic thermal flow field and microscopic molecular characteristics, effectively distinguishing gas components with similar thermal properties and significantly improving the selectivity and identification accuracy of the sensor in complex background gases.

[0066] Example 2

[0067] Please see Figure 3 This invention provides a gas thermal property parameter identification system based on MEMS sensors, and the technical solution is as follows:

[0068] Gas thermophysical property parameter identification system based on MEMS sensors, such as Figure 3 As shown, it includes:

[0069] Drive module: applies multi-frequency AC excitation current to the heating element of the MEMS sensor; causes the heating element to generate Joule thermal fluctuations at twice the reference frequency, resulting in the resistance value of the heating element fluctuating at the same frequency;

[0070] Extraction module: Real-time detection and extraction of the third harmonic output voltage signal generated across the heating element via a locked amplification circuit; decomposition of the third harmonic output voltage signal into an in-phase component signal that is in phase with the AC excitation current and an orthogonal component signal that has a phase deviation from the AC excitation current using phase-sensitive detection technology; calculation of the complex temperature rise of the heating element based on the in-phase component signal and the orthogonal component signal; and construction of a complex thermal impedance model of the gas under test based on the Joule thermal fluctuations.

[0071] Calculation module: Calculates the thermal conductivity of the gas under test based on the slope of the real part signal of the complex thermal impedance model as a function of the logarithm of the reference frequency; extracts the thermal diffusivity of the gas under test based on the imaginary part signal and phase delay of the complex thermal impedance model.

[0072] Analysis module: Utilizing the thermal conductivity and thermal diffusivity, combined with the correlation model, the volumetric specific heat capacity of the gas under test is decoupled and output to quantitatively analyze the component ratio of the mixed gas.

[0073] The gas thermophysical parameter identification system based on MEMS sensors has demonstrated high integration and stability in practical industrial environment monitoring scenarios. The core architecture of this identification system consists of multiple finely crafted modules that work together to complete a closed-loop processing procedure from the excitation of the underlying physical signal to the decoupling of the high-level thermophysical parameters.

[0074] The system's drive module, serving as the starting point of the entire identification process, is responsible for generating high-fidelity multi-frequency AC excitation current. In actual operation, the digital frequency synthesizer first generates multiple sets of reference sine wave signals with different preset frequencies in the digital domain. To address noise interference in specific frequency bands under complex environments, the system utilizes linear weighted superposition technology to synthesize these signals into a composite signal containing multiple target reference frequency components. Through closed-loop control of the constant current drive circuit, this composite current is precisely delivered to the heating element of the microelectromechanical system (MEMS) sensor. At this point, the system dynamically adjusts the weights of each reference frequency based on the acquired real-time thermal response feedback. This dynamic adjustment mechanism effectively compensates for low-frequency noise caused by environmental vibrations or parasitic capacitance effects at high frequencies. By enhancing the gain weight of specific frequencies or adjusting phase compensation parameters, it ensures that the heating element generates a highly linear thermal response within a very small temperature rise range, laying the foundation for subsequent high signal-to-noise ratio extraction.

[0075] When a multi-frequency AC excitation current flows through the heating element, a deep electrothermal modulation process occurs within the system. The heating element absorbs Joule heat, generating temperature fluctuations at twice the reference frequency. Based on the material's temperature coefficient of resistance, this temperature field change is linearly mapped to dynamic fluctuations in resistance. According to the nonlinear effect of Ohm's law, the fundamental frequency current interacts with the second-harmonic frequency resistance fluctuations, generating a third-harmonic frequency output voltage signal across the heating element that carries the "fingerprint" of the gas's thermal properties. To ensure the accuracy of this modulation process, the system strictly controls the operating temperature rise of the heating element within a linear range of 1K to 20K, and the highest reference frequency is always lower than the system's thermal relaxation frequency, thus allowing Joule heat energy to be instantaneously transferred to the gas medium, establishing a quasi-steady-state thermal wave diffusion field.

[0076] The system's extraction module captures the aforementioned weak signal using a high-performance lock-in amplifier circuit. This module employs a parallel quadrature demodulation architecture, utilizing multiple independent digital multipliers to decompose the third harmonic voltage signal into in-phase and quadrature components. To eliminate intermodulation interference between multi-frequency signals, the system uses an adaptive low-pass filter with a cutoff frequency less than half the minimum frequency difference between adjacent frequencies, filtering out harmonic noise while retaining pure real and imaginary characteristic data. Subsequently, the system performs magnitude normalization and gain compensation based on pre-calibrated amplitude-frequency response curves, eliminating the nonlinear attenuation of analog circuits over a wide bandwidth and ensuring the accuracy of characteristic component calculations.

[0077] After acquiring accurate signal components, the system enters the computation module for physical modeling and parameter inversion. The system first calculates the complex temperature rise and phase hysteresis using in-phase and quadrature components, and then constructs a complex thermal impedance model by combining this with the generated Joule thermal fluctuation power. To eliminate interference from non-gaseous factors, the system uses the substrate parasitic thermal impedance, pre-measured in a vacuum environment, and eliminates heat loss conducted through the sensor substrate via complex parallel cancellation operations. Simultaneously, geometric factor correction coefficients, including parameters such as linewidth and line length, and transmission line phase compensation are introduced to ensure that the constructed model accurately represents the heat transfer resistance and thermal capacity response within the gas.

[0078] Next, the system extracts the slope of the linear mapping between the real part of the complex thermal impedance model and the logarithm of the frequency, and obtains the thermal conductivity of the gas under test by inversion based on the analytical solution of the infinitely long linear heat source model. For the extraction of thermal diffusivity, the system uses the functional characteristics of the imaginary part of the signal for intercept fitting, sets an appropriate lower frequency limit to eliminate interference from chamber boundary reflections, and sets an upper frequency limit to satisfy the linear thermal conduction condition. This method based on frequency boundary locking and analytical model inversion eliminates the interference of contact thermal resistance on the phase signal, ensures the physical rigor of the extracted thermophysical parameters, and provides a standardized process for rapid parameter inversion under complex operating conditions.

[0079] Finally, the system's analysis module constructs a feature matrix of thermal parameters based on the physical correlation formulas of thermal diffusivity, thermal conductivity, and volumetric specific heat capacity. To prevent error accumulation caused by direct calculations, the system employs weighted least squares for consistency checks, and iterative optimization minimizes the residuals of each parameter under the correlation formulas. Based on the decoupled high-precision volumetric specific heat capacity, the system constructs a three-dimensional feature vector composed of three thermal dimensions and inputs it into the component mapping model. By searching for target positions in the feature matrix space and calculating offsets, the system ultimately achieves accurate quantitative inversion of the concentrations of each component in the mixed gas.

[0080] This system achieves precise measurement of gas thermal properties through a multi-module collaborative frequency domain identification architecture. The drive and extraction modules work together, utilizing third-harmonic generation detection technology to effectively separate the weak thermal response signal from the fundamental frequency electrical background, significantly improving the measurement signal-to-noise ratio. By constructing a complex thermal impedance model, the system can simultaneously decouple thermal conductivity, thermal diffusivity, and volumetric specific heat capacity, eliminating the degeneracy in group identification inherent in traditional single-parameter sensors. This physical mechanism-based parameter analysis method not only improves the distinguishability of gases with similar thermal conductivity but also provides data support for the real-time, quantitative inversion of mixed gas components.

[0081] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for identifying gas thermal property parameters based on MEMS sensors, characterized in that, include: Apply multi-frequency AC excitation current to the heating element of the MEMS sensor; This causes the heating element to generate Joule thermal fluctuations at twice the reference frequency, resulting in the resistance value of the heating element fluctuating at the same frequency. The triple-frequency output voltage signal generated at both ends of the heating element is detected and extracted in real time by locking the amplifier circuit; The third harmonic output voltage signal is decomposed into an in-phase component signal that is in phase with the AC excitation current and an orthogonal component signal that has a phase deviation from the AC excitation current using phase-sensitive detection technology. The complex temperature rise of the heating element is calculated based on the in-phase and orthogonal component signals, and a complex thermal impedance model of the gas under test is constructed by combining the Joule thermal fluctuations. The complex temperature rise amplitude of the heating element at twice the reference frequency and its phase lag relative to the AC excitation current are calculated using the in-phase, orthogonal component signals, and the temperature coefficient of resistance of the heating element. The Joule thermal fluctuation power generated by the heating element is calculated based on the effective value of the AC excitation current and the base resistance value of the heating element, and is defined as the dynamic heat flux vector input to the gas environment under test. The complex temperature rise amplitude and the dynamic heat flux vector are then divided by a complex number to obtain the complex thermal impedance model. The thermal conductivity of the gas under test is calculated based on the slope of the real part signal of the complex thermal impedance model as a function of the logarithm of the reference frequency. The thermal diffusivity of the gas under test is extracted based on the imaginary part signal and phase delay of the complex thermal impedance model. The steps for calculating thermal conductivity and thermal diffusivity include: extracting the real part signal of the complex thermal impedance model within a preset frequency range, establishing a linear mapping relationship between the real part signal and the reference frequency; calculating the slope of the linear mapping relationship, and inverting the thermal conductivity of the gas to be tested by combining the geometric characteristic parameters of the heating element; and using the imaginary part signal of the complex thermal impedance model, combined with the thermal conductivity, calculating and extracting the thermal diffusivity of the gas to be tested through intercept fitting. Using the thermal conductivity and thermal diffusivity, combined with a correlation model, the volumetric specific heat capacity of the gas to be tested is decoupled and output to quantitatively analyze the component ratio of the mixed gas. The step of outputting the volumetric specific heat capacity includes: establishing a thermal parameter characteristic matrix of the gas to be tested based on the physical correlation formula between thermal diffusivity, thermal conductivity, and volumetric specific heat capacity; the physical correlation formula satisfies the mathematical relationship that thermal diffusivity equals thermal conductivity divided by volumetric specific heat capacity; substituting the calculated thermal conductivity and the extracted thermal diffusivity into the physical correlation formula, the volumetric specific heat capacity of the gas to be tested is decoupled through numerical division.

2. The method for identifying gas thermal property parameters based on MEMS sensors according to claim 1, characterized in that: The steps of applying multi-frequency AC excitation current include: generating n sets of reference sine wave signals with different preset frequencies using a digital frequency synthesizer, and linearly weighting and superimposing the n sets of reference sine wave signals to synthesize a composite signal containing n target reference frequency components; inputting the composite signal to a constant current drive circuit, and outputting the multi-frequency AC excitation current to the heating element through closed-loop control; adjusting the weight of each reference frequency in the multi-frequency AC excitation current according to the thermal response feedback of the gas to be tested, so that the resistance value generates a same-frequency fluctuation containing n components twice the reference frequency, and generates a triple-frequency output voltage signal corresponding to each of the reference frequencies.

3. The method for identifying gas thermal property parameters based on MEMS sensors according to claim 1, characterized in that, The specific steps for generating the triplet output voltage signal include: the multi-frequency AC excitation current causes the heating element to generate temperature fluctuations containing twice the frequency components of each reference frequency, and the temperature fluctuations are linearly modulated into a dynamic resistance signal containing twice the reference frequency components based on the temperature coefficient of resistance; according to Ohm's law, the multi-frequency AC excitation current with a frequency of reference frequency and the dynamic resistance signal with a frequency of twice the reference frequency interact at both ends of the heating element to generate an original voltage signal containing a component of three times the reference frequency; the original voltage signal is processed by the lock-in amplifier circuit to filter out the reference frequency components and extract the triplet output voltage signal characterizing the thermophysical properties of the gas under test.

4. The method for identifying gas thermal property parameters based on MEMS sensors according to claim 1, characterized in that, The specific steps for extracting and decomposing the in-phase and quadrature component signals include: using a digital frequency synthesizer to generate a first reference sine signal that is in phase and has the same frequency as three times the reference frequency, and a second reference sine signal that has a 90° phase difference with the first reference sine signal; using phase-sensitive detection technology, multiplying the third harmonic output voltage signal with the first and second reference sine signals respectively, modulating the target frequency signal to the zero frequency band, and generating a mixed signal containing twice the three times the reference frequency; using a low-pass filter to filter the mixed signal, filtering out high-frequency components, and extracting in real time the in-phase component signal representing the real part of the third harmonic output voltage signal, and the quadrature component signal representing the imaginary part of the output voltage signal.

5. A gas thermal property parameter identification system based on MEMS sensors, characterized in that, include: Drive module: applies multi-frequency AC excitation current to the heating element of the MEMS sensor; This causes the heating element to generate Joule thermal fluctuations at twice the reference frequency, resulting in the resistance value of the heating element fluctuating at the same frequency. Extraction module: Real-time detection and extraction of the triple-frequency output voltage signal generated at both ends of the heating element through a locked amplification circuit; The third harmonic output voltage signal is decomposed into an in-phase component signal that is in phase with the AC excitation current and an orthogonal component signal that has a phase deviation from the AC excitation current using phase-sensitive detection technology. The complex temperature rise of the heating element is calculated based on the in-phase and orthogonal component signals, and a complex thermal impedance model of the gas under test is constructed by combining the Joule thermal fluctuations. The complex temperature rise amplitude of the heating element at twice the reference frequency and its phase lag relative to the AC excitation current are calculated using the in-phase, orthogonal component signals, and the temperature coefficient of resistance of the heating element. The Joule thermal fluctuation power generated by the heating element is calculated based on the effective value of the AC excitation current and the base resistance value of the heating element, and is defined as the dynamic heat flux vector input to the gas environment under test. The complex temperature rise amplitude and the dynamic heat flux vector are then divided by a complex number to obtain the complex thermal impedance model. Calculation module: Calculates the thermal conductivity of the gas under test based on the slope of the real part signal of the complex thermal impedance model as a function of the logarithm of the reference frequency; The thermal diffusivity of the gas under test is extracted based on the imaginary part signal and phase delay of the complex thermal impedance model. The steps for calculating thermal conductivity and thermal diffusivity include: extracting the real part signal of the complex thermal impedance model within a preset frequency range, establishing a linear mapping relationship between the real part signal and the reference frequency; calculating the slope of the linear mapping relationship, and inverting the thermal conductivity of the gas to be tested by combining the geometric characteristic parameters of the heating element; and using the imaginary part signal of the complex thermal impedance model, combined with the thermal conductivity, calculating and extracting the thermal diffusivity of the gas to be tested through intercept fitting. Analysis module: Utilizing the thermal conductivity and thermal diffusivity, combined with a correlation model, the volumetric specific heat capacity of the gas under test is decoupled and output to quantitatively analyze the component ratio of the mixed gas. The step of outputting the volumetric specific heat capacity includes: establishing a thermal parameter characteristic matrix of the gas under test based on the physical correlation formula between thermal diffusivity, thermal conductivity, and volumetric specific heat capacity; the physical correlation formula satisfies the mathematical relationship that thermal diffusivity equals thermal conductivity divided by volumetric specific heat capacity; substituting the calculated thermal conductivity and the extracted thermal diffusivity into the physical correlation formula, the volumetric specific heat capacity of the gas under test is decoupled through numerical division.

Images