MEMS thermal flow sensor system based on three-element differential structure

Abstract

The invention relates to the technical field of MEMS (micro electro mechanical system) sensing, in particular to an MEMS (micro electro mechanical system) thermal flow sensor system based on a three-element differential structure, which comprises a sensor module comprising an upstream temperature-sensitive element, a central heating element and a downstream temperature-sensitive element, the upstream temperature-sensitive element and the downstream temperature-sensitive element are arranged in geometric symmetry and thermal symmetry relative to the central heating element; the signal processing unit is used for collecting temperature signals of the upstream temperature-sensitive element and the downstream temperature-sensitive element and calculating differential temperature signals of the upstream temperature-sensitive element and the downstream temperature-sensitive element; wherein the signal processing unit is configured to identify the flow direction of the fluid based on the differential temperature signal, calculate an initial flow value, and correct the initial flow value based on the real-time working parameters and a preset compensation model so as to output a compensated flow measurement value; through sensor topology, multi-dimensional information fusion and an intelligent compensation algorithm, environment temperature excursion can be effectively restrained, different measurement media can be adapted, and the precision, stability and applicability of flow measurement are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of microelectromechanical systems (MEMS) sensing technology, and more particularly to a MEMS thermal flow sensor system based on a three-element differential structure. Background Technology

[0002] Thermal flow sensors are based on the principle of convection heat transfer. They infer flow rate by measuring the temperature field change caused by fluid flowing through a heating element. They have advantages such as high sensitivity, fast response, and no moving parts, and are widely used in industrial process control, environmental monitoring, medical devices and other fields.

[0003] Traditional thermal flow sensors typically employ a differential structure with dual upstream and downstream temperature-sensitive elements, detecting flow rate by measuring the temperature difference between the upstream and downstream of the heating element. However, such sensors face several key technical challenges in practical applications: First, changes in ambient temperature directly affect the reference resistance value of the sensitive element and the thermal properties of the medium, causing drift in the sensor's zero point (output at zero flow) and sensitivity (the relationship between output and flow rate), severely reducing measurement accuracy and long-term stability. Second, when the type of the measured medium changes or its thermal properties (such as specific heat capacity and thermal conductivity) fluctuate, the sensor's output characteristics change accordingly, leading to poor consistency in cross-medium measurements. Furthermore, the sensor chip's own thermal design, such as the thermal inertia of the heating element and the thermal diffusion effect of the substrate, affects the establishment speed and distribution symmetry of the temperature field, thereby limiting the sensor's dynamic response performance and measurement linearity.

[0004] In existing technologies, to solve the temperature drift problem, it is common to add an additional independent temperature sensor to monitor the ambient temperature for one-dimensional compensation, or to use a complex constant temperature difference control circuit. However, these methods often have a single compensation dimension, making it difficult to simultaneously correct the nonlinear drift caused by changes in both ambient temperature and heating power, and they are not adaptable enough to changes in medium properties. In addition, the traditional dual-element structure still has limitations in suppressing common-mode interference (such as overall changes in ambient temperature) and improving signal symmetry.

[0005] Therefore, to address the above problems, a MEMS thermal flow sensor system based on a three-element differential structure is proposed. Through sensor topology, multi-dimensional information fusion and intelligent compensation algorithm, it can effectively suppress environmental temperature drift, adapt to different measurement media, and significantly improve the accuracy, stability and applicability of flow measurement. Summary of the Invention

[0006] To overcome the problems that traditional thermal flow sensors often have a single compensation dimension when dealing with temperature drift, making it difficult to simultaneously correct nonlinear drift caused by changes in both ambient temperature and heating power, and lacking adaptability to changes in media properties.

[0007] The technical solution of this invention is: a MEMS thermal flow sensor system based on a three-element differential structure, comprising: The sensor module is disposed in a fluid microchannel and includes an upstream temperature-sensitive element, a central heating element and a downstream temperature-sensitive element arranged sequentially along the fluid flow direction. The upstream and downstream temperature-sensitive elements are arranged in a geometrically and thermally symmetrical manner relative to the central heating element. The signal processing unit is electrically connected to the sensor module and is used to acquire the temperature signals of the upstream and downstream temperature-sensitive elements, calculate their differential temperature signals, and acquire the real-time operating parameters of the central heating element. The signal processing unit is configured to: identify the fluid flow direction based on the differential temperature signal and calculate the initial flow rate value, and correct the initial flow rate value based on the real-time operating parameters and the preset compensation model, so as to output the compensated flow rate measurement value.

[0008] Preferably, the central heating element is a thin-film resistor structure disposed on a microchannel substrate, and a heat-insulating microstructure is disposed between it and the microchannel substrate. The upstream and downstream temperature-sensitive elements are both thin-film resistors made of the same material as the central heating element.

[0009] Preferably, the signal processing unit includes a temperature drift self-compensation module, which is configured to perform the following steps: The current ambient temperature, the current heating power of the central heating element, and the differential temperature signal are obtained from the upstream and downstream temperature-sensitive elements. Using the current ambient temperature and current heating power as input parameters, a preset multidimensional fitting model is invoked to calculate the sensitivity compensation coefficient and zero-point offset for the current operating condition. The initial flow rate value obtained based on the differential temperature signal is calculated and corrected in real time using the sensitivity compensation coefficient and the zero-point offset.

[0010] Preferably, the preset multidimensional fitting model is established in the following way: Under multiple different constant ambient temperatures, the central heating element is controlled to operate at multiple different constant power levels, and the zero-point value of the differential temperature signal is recorded under zero flow conditions, while the response value of the differential temperature signal is recorded under multiple known standard flow rates. Based on the dataset of ambient temperature, heating power, zero point value, and response value, a first fitting function for the zero point offset and a second fitting function for the sensitivity coefficient are established using a surface fitting method.

[0011] Preferably, the signal processing unit further includes a medium characteristic identification module, which is configured as follows: Under constant heating power and zero flow conditions, the current fluid medium type or thermophysical parameter range is identified based on the temperature change rate or steady-state temperature value of the upstream and downstream temperature-sensitive elements. The parameters in the preset compensation model are switched or adjusted based on the identification results.

[0012] Preferably, the system further includes a drive control unit connected to the central heating element, which is used to adjust the operating mode of the central heating element according to the instructions of the signal processing unit or external settings. The operating mode includes at least a constant power mode and a constant temperature difference mode.

[0013] Preferably, in constant temperature difference mode, the drive control unit is configured as follows: With the goal of maintaining a constant temperature difference between the average temperature of the upstream and downstream temperature-sensitive elements and the temperature of the central heating element, the heating power of the central heating element is dynamically adjusted. The change in heating power during the adjustment process is used as the main basis for calculating fluid flow rate.

[0014] Preferably, the upstream and downstream temperature-sensitive elements also serve as auxiliary heating elements, and the system is configured to detect the consistency of their thermoelectric characteristics by applying a brief excitation current to the upstream and / or downstream temperature-sensitive elements during startup or self-test phases.

[0015] Preferably, the signal processing unit is configured to synchronously sample the original temperature signals of the upstream and downstream temperature-sensitive elements, and then use a digital filtering algorithm to suppress common-mode noise before calculating the differential temperature signal.

[0016] Preferably, the sensor module, signal processing unit, and drive control unit are integrated into a single MEMS chip or system-in-package.

[0017] The beneficial effects of this invention are: The system of this invention establishes an excellent differential detection foundation through a symmetrical three-element topology; by fusing multi-dimensional information from ambient temperature, heating power, and differential signals, a high-precision real-time temperature drift compensation model is established; further, by combining intelligent functions such as media identification, operating mode switching, and component self-testing, a highly adaptive flow measurement solution is formed; specifically, this invention fundamentally improves the zero-point stability and sensitivity consistency of the sensor under various ambient temperatures and operating points, achieving accurate measurement over a wide temperature range; moreover, the system possesses preliminary media identification and adaptive capabilities, reducing measurement errors caused by changes in the measured medium and expanding the application range; simultaneously, through optimized thermal structure design and multi-mode driving, dynamic response and linearity are improved. Attached Figure Description

[0018] Figure 1 The diagram shown is a schematic of the overall architecture of the MEMS thermal flow sensor system based on a three-element differential structure according to the present invention. Figure 2 The diagram shown is a schematic of the metal thin-film resistor structure of the MEMS thermal flow sensor system based on a three-element differential structure according to the present invention. Figure 3 The diagram shown is a schematic of the temperature drift self-compensation algorithm flow of the MEMS thermal flow sensor system based on a three-element differential structure according to the present invention. Figure 4 The diagram shown illustrates the constant temperature difference mode control process of the MEMS thermal flow sensor system based on a three-element differential structure according to the present invention. Detailed Implementation

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0020] Example 1 Please see Figure 1 , Figure 2 , Figure 3 and Figure 4 The present invention provides an embodiment: a MEMS thermal flow sensor system based on a three-element differential structure, including a sensor module and a signal processing unit; The sensor module is set in the microchannel through which the fluid flows. Its core is a three-element combination consisting of an upstream temperature-sensitive element, a central heating element, and a downstream temperature-sensitive element. The upstream and downstream temperature-sensitive elements are symmetrically arranged on both sides of the central heating element. This symmetry is not only reflected in the geometric position but also in the matching of thermal characteristics, which is used to construct an ideal thermal dipole. The signal processing unit is electrically connected to the sensor module. It synchronously acquires temperature signals (typically expressed as resistance or voltage values) sensed by upstream and downstream temperature-sensitive elements and calculates the difference between them in real time, i.e., the differential temperature signal (ΔT). Simultaneously, the signal processing unit also acquires real-time operating parameters of the central heating element, such as applied voltage, current, or calculated real-time heating power (P). The signal processing unit has a pre-stored or computable compensation model. Its working logic is as follows: First, it uses the polarity (positive or negative) of the differential temperature signal ΔT to directly and quickly determine the fluid flow direction (from upstream to downstream or vice versa). The system calculates an initial flow rate value (Q_raw) based on the amplitude of ΔT. However, this initial value Q_raw is affected by the ambient temperature (T_env) and the current heating state (P). Therefore, the signal processing unit further takes the acquired real-time ambient temperature T_env (obtained from the average value of upstream and downstream temperature-sensitive elements) and real-time heating power P as inputs and calls the built-in compensation model. This model outputs a sensitivity correction coefficient (K_comp) and a zero-point offset (Offset_comp) for the current specific operating conditions. Finally, by performing an operation on the initial flow rate value in the form of "Q_final = K_comp * Q_raw + Offset_comp" (or other equivalent algorithms), a more accurate final flow rate measurement value Q_final is output after real-time correction. This process achieves source compensation of the flow rate signal.

[0021] Furthermore, the central heating element employs a low thermal inertia metal thin-film resistive structure (such as platinum, nickel, etc.) directly fabricated on the microchannel substrate. A specific thermal insulation microstructure is formed between the heating element and the substrate through etching or deposition processes, such as a suspended film, porous silicon, a low thermal conductivity material layer, or a microbridge structure. This thermal insulation microstructure can significantly reduce the dissipation of heat generated by the heating element towards the substrate, forcing more heat to be transferred through fluid convection and conduction, thereby improving heating efficiency, enhancing the asymmetry of the upstream and downstream temperature fields (i.e., differential signal), and reducing interference caused by substrate temperature fluctuations. Preferably, the upstream and downstream temperature-sensitive elements are fabricated using the same thin-film materials and processes as the central heating element to ensure a high degree of consistency in the thermoelectric properties (such as the temperature coefficient of resistance, TCR) of the three elements.

[0022] Furthermore, the signal processing unit integrates a dedicated temperature drift self-compensation module; the operation of this module follows the following process: Step S1: Parameter Synchronization Acquisition: Within the same sampling period, acquire the current average ambient temperature (T_env) measured and calculated by the upstream and downstream temperature-sensitive elements, acquire the current working heating power (P) of the central heating element obtained by feedback or calculation from the drive circuit, and acquire the upstream and downstream differential temperature signal (ΔT) calculated in real time. Step S2: Model Query and Calculation: The parameter pair (T_env, P) is used as an index or input variable and fed into a pre-calibrated and stored multidimensional fitting model. After training, this model outputs two key compensation parameters: one is the sensitivity compensation coefficient (S) for the current (T_env, P) state, used to correct the slope of the flow-ΔT curve; the other is the zero-point offset (O) for the current state, used to correct the ΔT baseline value at zero flow. Step S3: Real-time Correction Application: Using the obtained S and O, compensate for the original flow rate value calculated based on the current ΔT; for example, the formula can be used: Q_comp = S * f(ΔT) + O, where f(ΔT) is a function that converts ΔT into the original flow rate. Through this process, bivariate adaptive compensation for ambient temperature and heating conditions is achieved.

[0023] Furthermore, the pre-defined multidimensional fitting model is established in the following way: Calibration Phase: Place the sensor in a controllable ambient temperature chamber; at multiple different constant ambient temperature points (T_i, i=1,2,...,m) covering the expected operating range, control the central heating element to operate at multiple different constant power levels (P_j, j=1,2,...,n); for each (T_i, P_j) combination of operating conditions: a. Under the condition of ensuring zero flow, record the stable differential temperature signal value at this time, that is, the "zero point value" Z_ij under this operating condition; b. Introduce a series of known and precise standard flow rates (Q_k, k=1,2,...,p) and record the corresponding stable differential temperature signal value ΔT_ijk at each standard flow rate; Modeling Phase: Collect all calibration data; for zero-point data, with (T, P) as independent variables and the zero-point value Z as dependent variable, establish the "first fitting function" describing the zero-point drift law through bivariate surface fitting (such as polynomial regression, spline interpolation, etc.): Z = F_z(T, P); for sensitivity data, under each (T_i, P_j) condition, a flow-ΔT relationship curve can be fitted using the (Q_k, ΔT_ijk) data, and its slope is the sensitivity S_ij under that condition; then, with (T, P) as independent variables and S as dependent variable, establish the "second fitting function" through surface fitting: S = F_s(T, P); the functions F_z and F_s together constitute a multidimensional model for real-time compensation.

[0024] Furthermore, the system also includes a medium characteristic identification module, which uses the sensor's own thermal response characteristics to make a preliminary judgment on the fluid medium; During system startup, intermittent periods, or specific self-test commands, the central heating element operates at a fixed power P_test, temporarily maintaining fluid stillness (zero flow). At this time, the temperature rise process or final steady-state temperature value of the upstream and downstream temperature-sensitive elements (which are also temperature sensors) strongly depends on the thermal properties (such as thermal diffusivity) of the surrounding fluid. The module monitors the temperature change curves (transient response) of the two temperature-sensitive elements over time or records the achieved steady-state temperature value (steady-state response). Different media (such as air, nitrogen, helium, and water) have distinctly different thermal properties, thus exhibiting characteristic temperature rise curves or steady-state values. The module compares the measured response pattern with a pre-stored database of various typical media characteristics or extracts characteristic parameters (such as the temperature rise time constant) through algorithms to identify the current fluid's medium type or its thermal property parameter range. Based on the identification results, the signal processing unit can automatically switch to a specific compensation model parameter set optimized for that type of medium, or dynamically adjust the coefficients of the general compensation model, thereby achieving measurement capabilities adaptable to various media and improving the system's versatility.

[0025] Furthermore, the system also includes a drive control unit connected to the central heating element. This unit receives instructions from the signal processing unit or an external system and can support at least two operating modes: constant power mode (CP) and constant temperature difference mode (CTD). In constant power mode, the drive control unit maintains a constant power of the heating element and measures the change in the upstream and downstream temperature difference ΔT. In constant temperature difference mode, the drive control unit receives instructions from the signal processing unit to maintain a preset constant temperature difference (e.g., maintaining a constant temperature difference between the central heating element and the average upstream and downstream temperatures) and dynamically and rapidly adjusts the power P applied to the central heating element through a closed-loop control algorithm (such as PID).

[0026] Furthermore, in Constant Temperature Difference (CTD) mode, the drive control unit executes the following control logic: a target temperature difference ΔT_target is set (e.g., the heating element is at a certain temperature higher than the fluid environment); the signal processing unit calculates the actual temperature difference ΔT_actual between the temperature of the central heating element and the average temperature of the upstream and downstream temperature-sensitive elements in real time; the drive control unit compares ΔT_actual with ΔT_target, generates an error signal, and adjusts the heating power P accordingly to make ΔT_actual track ΔT_target; when the fluid flow rate changes, convective heat dissipation increases or decreases, and in order to maintain a constant ΔT_actual, the drive control unit must increase or decrease the heating power P accordingly; therefore, in this mode, the change in heating power P itself becomes a direct, highly linear signal reflecting the flow rate; this mode is suitable for measurement scenarios with low flow rates and high dynamic response.

[0027] Furthermore, upstream and downstream temperature-sensitive elements can be reused as auxiliary heating elements at specific stages. For example, during the power-on initialization, periodic self-calibration, or fault diagnosis phases of the sensor system, the signal processing unit can control a switching circuit to temporarily apply a known, short-term excitation current (much smaller than the normal heating current) to the upstream and / or downstream temperature-sensitive elements. By measuring the voltage drop or temperature change of the element itself at this time, its resistance value and temperature coefficient of resistance (TCR) at the current temperature can be accurately calculated. By comparing these thermoelectric parameters of the two elements, the consistency of their characteristics can be evaluated in real time. If the inconsistency exceeds the tolerance, an additional correction factor can be introduced in the signal processing, or an alarm can be triggered in case of a fault, thereby improving the reliability and long-term stability of the system.

[0028] Furthermore, to improve the signal-to-noise ratio and common-mode rejection ratio, the signal processing unit employs high-precision synchronous sampling technology for signal acquisition from upstream and downstream temperature-sensitive elements, ensuring that the two signals are captured at the same time to eliminate errors introduced by timing differences. The acquired raw signals are amplified and then sent to the digital processing core (such as MCU or DSP). Before or after calculating the differential signal, digital filtering algorithms (such as moving average or low-pass filtering) are used to process the signal, focusing on suppressing common-mode noise caused by slow changes in ambient temperature, thereby further highlighting the flow-related differential signal components.

[0029] Furthermore, the sensor module, signal processing unit (including ADC, processor, and memory), and drive control unit (including power drive circuit) can be integrated into a single MEMS chip using advanced integration technology or integrated into a compact package using system-in-package (SiP) technology. This integrated structure reduces external leads and connections, lowers noise interference and parasitic effects, and improves the overall reliability, miniaturization, and portability of the system, making it suitable for embedded devices or portable instruments with strict space constraints.

[0030] Through the above steps, the system of this invention constructs an excellent differential detection foundation through a symmetrical three-element topology; by integrating multi-dimensional information such as ambient temperature, heating power, and differential signals, a high-precision real-time temperature drift compensation model is established; further, by combining intelligent functions such as media identification, working mode switching, and component self-testing, a highly adaptive flow measurement solution is formed; specifically, this invention fundamentally improves the zero-point stability and sensitivity consistency of the sensor under various ambient temperatures and operating points, achieving accurate measurement over a wide temperature range; moreover, the system possesses preliminary media identification and adaptive capabilities, reducing measurement errors caused by changes in the measured medium and expanding the application range; simultaneously, through optimized thermal structure design and multi-mode driving, dynamic response and linearity are improved.

[0031] Example 2 Optionally, this embodiment provides the most basic implementation of the present invention.

[0032] The MEMS thermal flow sensor system based on a three-element differential structure in this embodiment includes a sensor module and a signal processing unit at its core.

[0033] The sensor module is fabricated on a silicon-based or other substrate suitable for MEMS processes. A microchannel for fluid passage is formed on the substrate surface using microfabrication techniques (such as photolithography, thin film deposition, and etching). In the central region of the microchannel, three key components are sequentially integrated along its length (i.e., the predetermined fluid flow direction): an upstream temperature-sensitive element R_us, a central heating element R_h, and a downstream temperature-sensitive element R_ds. All three components are made of a thin metal film (such as platinum Pt) with a high temperature coefficient of resistance (TCR), forming a thin-film resistor. The upstream temperature-sensitive element R_us and the downstream temperature-sensitive element R_ds are arranged strictly symmetrically in terms of geometry, size, and distance from the central heating element R_h.

[0034] The signal processing unit can be a microcontroller (MCU) or an application-specific integrated circuit (ASIC); its hardware includes: Signal conditioning and acquisition circuit: includes a multiplexer, instrumentation amplifier, and analog-to-digital converter (ADC). This circuit is connected to R_us, R_ds, and R_h. Processing core: such as the ARM Cortex-M series core, responsible for running control and compensation algorithms; Memory: Used to store calibration parameters, compensation models, and programs.

[0035] The system works as follows: Signal acquisition: The signal conditioning and acquisition circuit measures the resistance values ​​of R_us and R_ds in a synchronous sampling manner (achieved by applying a small constant current to measure the voltage), and converts them into the corresponding temperature values ​​T_us and T_ds; at the same time, by measuring the voltage V_h across R_h and the current I_h flowing through it, the real-time heating power P = V_h * I_h is calculated.

[0036] Calculate the differential signal and initial flow rate: The core processing calculates the upstream and downstream temperature difference ΔT = T_ds - T_us; if ΔT>0, it indicates that the fluid flows from upstream to downstream; if ΔT<0, it indicates reverse flow; according to a preset basic conversion function calibrated under standard operating conditions, such as a polynomial Q_raw = f(ΔT), the absolute value of ΔT is converted into the initial flow rate value Q_raw; this function is usually calibrated at a fixed ambient temperature T0 and heating power P0.

[0037] Obtain ambient temperature: The processing core calculates the current ambient (fluid) temperature reference value T_env = (T_us + T_ds) / 2.

[0038] Compensation calculation: The processing core takes the currently measured T_env and P as input and queries the compensation model stored in memory; the model is a two-dimensional lookup table or two binary functions, which output the sensitivity compensation coefficient K_comp and the zero offset Offset_comp at the current (T_env, P) respectively.

[0039] Output the final result: The initial flow is corrected using compensation parameters to obtain the final flow value Q_final = K_comp * Q_raw + Offset_comp; this value is output through a communication interface (such as I2C, SPI).

[0040] Example 3 Optionally, this embodiment optimizes the physical structure of the sensor module based on embodiment 2.

[0041] In this embodiment, the central heating element R_h is not directly fabricated on the substrate. First, a layer of silicon dioxide or silicon nitride is grown on the substrate as a sacrificial / insulating layer through thermal oxidation or LPCVD. Then, a platinum thin film pattern of R_h is fabricated on this layer. Afterward, a portion of the substrate material directly below R_h is selectively removed by anisotropic dry etching (such as DRIE) or wet etching to form a cavity or release a suspended thin film structure, thereby constituting a thermally insulating microstructure. This results in an insulating or low thermal conductivity support below R_h, significantly reducing the longitudinal conduction of heat to the substrate.

[0042] Similarly, the upstream temperature-sensitive element R_us and the downstream temperature-sensitive element R_ds are also fabricated on similar suspended microbridges or thin films. However, in order to make them more sensitive to fluid temperature, the thermal insulation structure below them may be designed to be thinner or have slightly higher thermal conductivity than that of R_h. The film materials, thicknesses and process steps of the three are exactly the same to ensure a high degree of consistency in TCR, which is the physical basis for achieving accurate differential measurement. This structure improves the efficiency of heat transfer of the heating element to the fluid, enhances the asymmetry of the upstream and downstream temperature fields, and suppresses thermal interference from the substrate.

[0043] Example 4 Optionally, this embodiment, based on embodiment 2 or 3, elaborates in detail the method for establishing the multidimensional temperature drift compensation model and the implementation of the medium identification function.

[0044] The establishment (calibration process) of the multidimensional compensation model is as follows: Step 1: Place the assembled sensor system in a high-precision ambient temperature test chamber and connect it to a standard flow generator and a precision data acquisition system; Step 2: Set the test chamber temperature to the first calibration point T1 (e.g., -10°C); after the temperature stabilizes, set the power of the central heating element to P1 (e.g., 10mW); under zero flow conditions, record the stable upstream and downstream temperature difference signal value, denoted as "zero point" Z11; then, sequentially introduce multiple known standard flow rates Qk (e.g., 0, 10, 20, ... sccm), and record the stable ΔT value ΔT11k at each flow rate; Step 3: Keep the ambient temperature T1 constant, change the heating power to P2 (e.g., 15mW), and repeat step 2 to obtain the zero point Z12 and the dataset ΔT12k; Step 4: Change the ambient temperature to T2, T3...Tm, and repeat steps 2 to 3 for each temperature point to finally obtain a complete data cube: for each (Ti, Pj) combination, there is a corresponding zero point Zij and a set of (ΔTijk,Qk) data; Step 5: Zero-point drift model fitting: Organize all zero-point data Zij into a dataset with T and P as independent variables; perform bivariate polynomial surface fitting using the least squares method; for example, using a second-order model: Z = a0 + a1*T + a2*P + a3*T*P + a4*T^2 + a5*P^2 Where Z represents the zero-point offset (the unit can be μV or directly ΔT value), T represents the ambient temperature (°C), P represents the heating power (mW), and a0~a5 are the fitting coefficients; the coefficients a0~a5 are stored in the memory.

[0045] B. Real-time compensation and media identification workflow: After the system is powered on, the temperature drift self-compensation module in the signal processing unit executes the following loop: Step 1: Synchronously collect T_us, T_ds, V_h, I_h; Step 2: Calculate T_env = (T_us+T_ds) / 2, P = V_h*I_h, ΔT = T_ds - T_us; Step 3: Medium Identification Step: During the first measurement cycle after system startup, or during periodic self-tests, the control drive circuit makes the central heating element R_h operate at a fixed identification power P_iden for a short time t_iden (e.g., 100ms), while maintaining the flow rate at zero or extremely low; process the temperature rise curves of core monitoring T_us and T_ds within the time t_iden; calculate the average temperature rise rate dT / dt, or the time constant τ required to reach a certain threshold temperature; there are significant differences in dT / dt or τ for different media (e.g., air and carbon dioxide); compare the calculated feature values ​​with the pre-stored "medium feature-compensation parameter mapping table" in memory to determine the most likely medium type, and select or adjust the model coefficient set used in subsequent compensation calculations (e.g., select a0~a5 and c0~c5 corresponding to "air"). Step 4: Compensation Calculation: Substitute the current T_env and P into the selected zero-point model and sensitivity model to calculate the current Offset_comp(T_env, P) and K_comp(T_env, P); Step 5: Calculate the raw flow rate based on the basic transformation function Q_raw = f(ΔT); Step 6: Calculate the final flow: Q_final = K_comp(T_env, P) * Q_raw + Offset_comp(T_env, P).

[0046] Example 5 Optionally, this embodiment describes a specific implementation of the system in constant temperature difference mode, based on any one of embodiments 2-4.

[0047] The system includes a programmable drive control unit (which may be a functional module within the signal processing unit) that can switch between constant power (CP) mode and constant temperature difference (CTD) mode according to instructions.

[0048] In CTD mode, the goal is to maintain a constant difference ΔT_target between the central heating element temperature T_h and the fluid ambient temperature T_env; the implementation steps are as follows: Temperature acquisition: R_h itself is also a resistor, and its temperature T_h can be calculated by measuring its resistance value when it is working; at the same time, T_env is calculated from R_us and R_ds; Error Calculation: The core processing calculates the actual temperature difference ΔT_actual = T_h - T_env, and calculates the error e = ΔT_target - ΔT_actual. ΔT_target is the target value set by the user or system (e.g., 30°C). Closed-loop control: The drive control unit adopts a proportional-integral (PI) control algorithm; the control quantity u(t) is the heating power P(t) that needs to be applied to make ΔT_actual approach ΔT_target; the discrete form of the PI algorithm is: P[k] = P[k-1]+ Kp * (e[k] - e[k-1]) + Ki * e[k] Where P[k] is the heating power of the kth control cycle, e[k] is the temperature difference error of the kth cycle, and Kp and Ki are the proportional and integral coefficients; Power drive: The drive control unit controls a power drive circuit based on the calculated P[k] through pulse width modulation (PWM) or digital-to-analog conversion (DAC) to adjust the voltage or current applied to R_h, thereby precisely controlling its heat generation power; Flow output: In steady state, in order to maintain a constant ΔT_target, the heating power P must accurately compensate for the heat carried away by the fluid; therefore, the larger the flow rate, the larger the required steady-state power P_steady; at this time, the flow rate Q is directly calculated from the steady-state heating power P_steady through a calibration function Q = g(P_steady); this function is usually obtained through calibration in CTD mode and has excellent linearity.

[0049] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A MEMS thermal flow sensor system based on a three-element differential structure, characterized in that: include: The sensor module is disposed in a fluid microchannel and includes an upstream temperature-sensitive element, a central heating element and a downstream temperature-sensitive element arranged sequentially along the fluid flow direction. The upstream and downstream temperature-sensitive elements are arranged in a geometrically and thermally symmetrical manner relative to the central heating element. The signal processing unit is electrically connected to the sensor module and is used to acquire the temperature signals of the upstream and downstream temperature-sensitive elements, calculate their differential temperature signals, and acquire the real-time operating parameters of the central heating element. The signal processing unit is configured to: identify the fluid flow direction based on the differential temperature signal and calculate the initial flow rate value, and correct the initial flow rate value based on the real-time operating parameters and the preset compensation model, so as to output the compensated flow rate measurement value.

2. The MEMS thermal flow sensor system based on a three-element differential structure according to claim 1, characterized in that: The central heating element is a thin-film resistor structure disposed on a microchannel substrate, and a heat-insulating microstructure is disposed between it and the microchannel substrate. The upstream and downstream temperature-sensitive elements are both thin-film resistors made of the same material as the central heating element.

3. The MEMS thermal flow sensor system based on a three-element differential structure according to claim 2, characterized in that: The signal processing unit includes a temperature drift self-compensation module, which is configured to perform the following steps: The current ambient temperature, the current heating power of the central heating element, and the differential temperature signal are obtained from the upstream and downstream temperature-sensitive elements. Using the current ambient temperature and current heating power as input parameters, a preset multidimensional fitting model is invoked to calculate the sensitivity compensation coefficient and zero-point offset for the current operating condition. The initial flow rate value obtained based on the differential temperature signal is calculated and corrected in real time using the sensitivity compensation coefficient and the zero-point offset.

4. The MEMS thermal flow sensor system based on a three-element differential structure according to claim 3, characterized in that: The preset multidimensional fitting model is established in the following way: Under multiple different constant ambient temperatures, the central heating element is controlled to operate at multiple different constant power levels, and the zero-point value of the differential temperature signal is recorded under zero flow conditions, while the response value of the differential temperature signal is recorded under multiple known standard flow rates. Based on the dataset of ambient temperature, heating power, zero point value, and response value, a first fitting function for the zero point offset and a second fitting function for the sensitivity coefficient are established using a surface fitting method.

5. A MEMS thermal flow sensor system based on a three-element differential structure according to claim 1, characterized in that: The signal processing unit further includes a medium characteristic identification module, which is configured as follows: Under constant heating power and zero flow conditions, the current fluid medium type or thermophysical parameter range is identified based on the temperature change rate or steady-state temperature value of the upstream and downstream temperature-sensitive elements. The parameters in the preset compensation model are switched or adjusted based on the identification results.

6. A MEMS thermal flow sensor system based on a three-element differential structure according to claim 1, characterized in that: The system also includes a drive control unit connected to the central heating element, which is used to adjust the working mode of the central heating element according to the instructions of the signal processing unit or external settings. The working mode includes at least a constant power mode and a constant temperature difference mode.

7. A MEMS thermal flow sensor system based on a three-element differential structure according to claim 6, characterized in that: In constant temperature difference mode, the drive control unit is configured as follows: With the goal of maintaining a constant temperature difference between the average temperature of the upstream and downstream temperature-sensitive elements and the temperature of the central heating element, the heating power of the central heating element is dynamically adjusted. The change in heating power during the adjustment process is used as the main basis for calculating fluid flow rate.

8. A MEMS thermal flow sensor system based on a three-element differential structure according to claim 1, characterized in that: The upstream and downstream temperature-sensitive elements also serve as auxiliary heating elements. The system is configured to detect the consistency of their thermoelectric characteristics by applying a brief excitation current to the upstream and / or downstream temperature-sensitive elements during startup or self-test phases.

9. A MEMS thermal flow sensor system based on a three-element differential structure according to claim 1, characterized in that: The signal processing unit is configured to synchronously sample the original temperature signals of the upstream and downstream temperature-sensitive elements, and then use a digital filtering algorithm to suppress common-mode noise before calculating the differential temperature signal.

10. A MEMS thermal flow sensor system based on a three-element differential structure according to any one of claims 1-9, characterized in that: The sensor module, signal processing unit, and drive control unit are integrated into a single MEMS chip or system-in-package.

Images