Asymmetric multi-channel thermal flow sensor for integrated component detection and signal processing method

Abstract

The invention relates to an asymmetric multichannel thermal flow sensor for integrated component detection and a signal processing method, and belongs to the technical field of sensors. The problems that a traditional symmetrical single-channel thermal flow sensor is limited in measuring range, saturated in output, insufficient in measuring precision and single in measuring flow are solved. The sensor comprises a flow inlet cavity, a flow outlet cavity and a measuring pipeline composed of five parallel circular pipelines, the middle pipeline is a component measuring pipeline, and the four pipelines on the periphery are flow measuring pipelines. A probe is arranged in the middle of the component measuring pipeline, a heating resistance wire is arranged in the middle of the flow measuring pipeline, and temperature measuring resistance wires are asymmetrically distributed on two sides. The pipeline is made of aluminum, and fillets are arranged at the joints to optimize fluid flow. The heating resistance wire is a nickel chrome wire, the temperature measuring resistance wire is a platinum wire, the component measuring probe is made of a platinum film, and the surfaces of the platinum film and the platinum film are insulated. An improved Wheatstone bridge circuit is adopted for signal collection, and precision and protectiveness are improved.

Description

Technical Field

[0001] This invention belongs to the field of sensor technology, specifically relating to an asymmetric thermal flow sensor with integrated component detection and a signal processing method. Background Technology

[0002] Flow measurement technology plays an indispensable role in modern industrial production, energy management, and daily life. Among various flow detection methods, thermal distributed flow meters have become a research focus in the field of flow detection technology due to their advantages such as small size, rapid response, and high measurement accuracy. A conventional thermal distributed flow sensor typically consists of three thermal sensors: symmetrically placed temperature sensing resistors and a central heating resistor. When there is no fluid flow in the pipe, the temperature is uniformly distributed, and the temperature difference between the two temperature sensing resistors is zero. When fluid flows in the pipe, the fluid carries away the heat from the heating resistors, changing the temperature field at both ends and causing a temperature difference between the two temperature sensing resistors. Finally, the fluid flow rate is measured based on the experimentally fitted relationship between the temperature difference and the flow rate.

[0003] Most thermal distributed gas flow meters on the market currently employ a single-channel measurement method, capable of measuring only flow rate data and unable to measure the composition of mixed gases. This single-channel thermal distributed flow meter relies on a single sensor to measure gas flow, which may limit its measurement accuracy when gas composition varies, ambient temperature fluctuates, or pressure changes significantly. Furthermore, this measurement method is non-linear; once the flow rate reaches a certain value, the temperature difference between the sensing resistor and the flow rate ceases to change, the sensor output saturates, and the measurement range is limited. While reducing the channel diameter and improving convective heat transfer can increase sensor sensitivity, reducing the channel diameter also reduces the sensor's measurement range. Therefore, new sensor structures and detection methods are needed to expand the sensor's measurement range without compromising sensitivity and accuracy. Simultaneously, to meet the needs of some practical production scenarios (such as hydrogen and natural gas transmission and distribution, and combustion efficiency control), designing a high-precision flow meter with gas component detection has become a key technological focus in the current industrial measurement field. Summary of the Invention

[0004] The purpose of this invention is to solve the problems of limited range, output saturation, insufficient measurement accuracy, and single-flow measurement of existing sensors caused by the traditional symmetrical single-channel thermal flow sensor design.

[0005] To achieve the above objectives, the present invention employs the following technical means:

[0006] On one hand, the present invention provides an asymmetric multi-channel thermal flow sensor with integrated component detection, comprising a flow inlet chamber, a flow outlet chamber, a flow measurement pipe, and a component measurement pipe. The flow inlet chamber and the flow outlet chamber each have corresponding flow inlets and outlets. The flow inlet chamber and the flow outlet chamber are connected by the flow measurement pipe and the component measurement pipe. The flow measurement pipe consists of four parallel circular pipes, and the component measurement pipe is one circular pipe parallel to the flow measurement pipe.

[0007] Furthermore, the overall material of the flow sensor structure is aluminum, and the outer diameter of the flow measurement and component measurement circular pipe is 10.5 mm, the inner diameter is 10 mm, and the length is 180 mm.

[0008] Furthermore, the connection points between the flow inlet cavity and the flow measurement pipe and the component measurement pipe, as well as the connection points between the flow outlet cavity and the flow measurement pipe and the component measurement pipe, are all provided with rounded corners, such as... Figure 2 As shown.

[0009] The above technical solution provides rounded corners at both ends of the measuring pipe to facilitate fluid inflow and outflow.

[0010] Furthermore, the heating resistance wire is connected to the power supply via a relay, which is controlled by a microcontroller outputting a PWM signal.

[0011] Furthermore, a platinum thin film probe integrated with MEMS technology is provided in the center of the component measurement pipe. The probe is formed into a micron-scale sensitive structure through photolithography and etching technology, and its arrangement direction is parallel to the gas flow direction.

[0012] The above technical solution reduces the disturbance of the probe to the gas flow and improves the measurement accuracy at higher gas flow rates.

[0013] Furthermore, the probe is connected to a relay and controlled by a microcontroller to achieve transient heating by applying a constant power pulse of 0.5–2 seconds. Subsequently, the resistance change of the probe is detected at a certain frequency to obtain the temperature change curve of the probe.

[0014] The above technical solution allows for the measurement of the thermal conductivity of a gas, and further, the determination of its composition. The principle is as follows: In an infinitely large, static gas medium, the probe temperature rise ΔT satisfies the following relationship with time t:

[0015]

[0016] Where C is a constant, q is the pulse power, and k is the gas thermal conductivity. It can be seen that the slope of the temperature rise curve, d(ΔT) / d(Int), is inversely proportional to the thermal conductivity k. Essentially, the higher the gas thermal conductivity, the slower the probe temperature rises under the same heating power. Furthermore, the composition of the gas can be determined based on its thermal conductivity. The relationship between thermal conductivity and composition is as follows:

[0017] λ mix =∑(Xi·λ i )

[0018] Where, λ mix Let λ be the thermal conductivity of the gas mixture, Xi be the proportion of the i-th gas, and λ be the thermal conductivity of the mixture. i Let be the thermal conductivity of the i-th gas. If there are only two gases in the mixture, then ∑(Xi) = 100%. Therefore, by solving the two equations simultaneously, the composition of the gas mixture can be obtained from its thermal conductivity. It is evident that this method is only applicable to the measurement of gas composition in a mixture of two single gases. Since there is a relationship between probe resistance change and thermal conductivity, and a relationship between thermal conductivity and gas composition, we can deduce that there is also a relationship between probe resistance change and gas composition. This relationship can be fitted experimentally.

[0019] Furthermore, in the flow measurement pipeline, there are small sections of pipe with a slightly smaller outer diameter than other positions at the positions of the heating resistance wire and the upstream and downstream temperature measuring resistance wires. The upstream and downstream temperature measuring resistance wires are asymmetrically distributed relative to the heating resistance wire. The distance between the upstream resistance wire and the heating resistance wire is 40mm, the distance between the downstream resistance wire and the heating resistance wire is 60mm, and the distance between the upstream and downstream resistance wires and both ends of the parallel circular pipe is 30mm.

[0020] The above technical solution facilitates the installation and positioning of the temperature measuring resistance wire and the heating resistance wire, delays the saturation of the sensor output, and increases the range of the gas flow sensor.

[0021] Furthermore, the heating resistance wire is a nickel-chromium resistance wire with a diameter of 0.1 mm, and the temperature measuring resistance wire is a platinum resistance wire with a diameter of 0.05 mm. The surfaces of both the heating resistor and the temperature measuring resistor are insulated.

[0022] The above technical solution uses nickel-chromium resistance wire as the heating resistor, which has a high temperature coefficient of resistance and good high temperature resistance, and a fast heating speed. Platinum resistance wire is used as the temperature measuring resistor, which has good stability, very linear temperature characteristics, and high temperature measurement accuracy.

[0023] Furthermore, thermally conductive material is placed between the heating resistor and the temperature measuring resistor and the circular pipe, and during the sensor measurement process, thermal insulation cotton is used to cover the flow measuring pipe.

[0024] Furthermore, in the above measurement scheme, the temperature difference signal of the temperature measuring resistors at both ends of the measuring tube is acquired by a signal acquisition circuit, amplified by a differential signal amplification circuit, and then filtered by a filtering circuit to remove noise. Subsequently, the analog signal is converted into a digital signal by the A / D converter of the microcontroller and processed internally by the microcontroller. Finally, the flow signal is transmitted to the mobile terminal and PC by the WIFI module and serial communication respectively.

[0025] Furthermore, the signal acquisition circuit employs a modified Wheatstone bridge, such as... Figure 3 As shown. Due to the inherent limitations in accuracy and sensitivity of traditional Wheatstone bridges, the above solution proposes a signal acquisition circuit with protective features. It consists of a platinum resistance wire R... S1 R S2 Together with precision resistors R2, R3, R4, and R5, it forms the circuit. The voltage signal acquired by the improved circuit is as follows:

[0026]

[0027] When R S1 =R S2 =R2=R3=R4=R5=R0, and ΔR S1 =ΔR,ΔR S2 When R = -ΔR (fluid flows in, upstream resistance temperature decreases, resistance increases; downstream resistance temperature increases, resistance decreases), the above formula can be simplified to:

[0028]

[0029] Through the above technical solution, the parallel circuit composed of R2, R3, R4 and R5 can control the circuit resistance value relatively easily. Furthermore, if one resistor in a parallel circuit fails, the other resistor can still play a role, preventing the circuit from completely failing and thus protecting the circuit.

[0030] Furthermore, in the above measurement scheme, to maintain a constant temperature of the central heating resistance wire and prevent the fluid flow from carrying away a large amount of heat, thus affecting the measurement accuracy, the current through the heating resistance wire is measured by the microcontroller's A / D converter, and the resistance value of the heating resistance is calculated, ultimately obtaining the current temperature of the resistance. Then, the current resistance temperature and the desired temperature are compared using PID calculations, and a PWM signal is output to control the operation of the relay, thereby ensuring that the temperature of the central heating resistance remains constant.

[0031] On the other hand, this invention also provides a signal processing method. Due to interference from some unknown factors, the acquired signal data may contain some anomalous data with large deviations, affecting the accuracy of the acquisition results. Therefore, it is necessary to process the signal data. The specific steps are as follows:

[0032] (I) Preliminary data processing using the standard deviation method. In the measurement experiment, the test data follows a normal distribution. Approximately 68% of the data falls within one standard deviation of the mean, about 95% falls within two standard deviations, and 99.7% falls within three standard deviations. First, the mean (μ) and standard deviation (σ) of the collected data sets are calculated. Then, the threshold for outliers is set at three standard deviations. If the data falls outside three standard deviations, the corresponding data is removed, and the remaining data are retained.

[0033] Standard deviation is usually unknown, but experimental standard deviation can be used instead. In a set of n independent and repeated observations, the k-th observation X... k The difference between the arithmetic means of the test data is called the residual error V. k ,but:

[0034] Therefore, the experimental standard deviation s(X) k )for:

[0035]

[0036] (II) Further processing using box plots. Sort the data after the first step in ascending order to find the first quartile (Q1): the number at the 25th percentile; the median: the number at the 50th percentile; and the third quartile: the number at the 75th percentile. The difference between the third quartile and the first quartile is called the interquartile range (IQR), an important statistic for measuring the dispersion of the data. That is: IQR = Q3 - Q1

[0037] The upper edge of the data is: M s =Q1 + 1.5 × IQR

[0038] The bottom edge of the data is: M x =Q3-1.5×IQR

[0039] Data falling outside the upper and lower edges of the data set is considered outlier, and the corresponding data is removed, while the remaining data is retained.

[0040] (III) The final step of data processing is to use k-means clustering.

[0041] Step 1: Determine the number of clusters: Choose a relatively small value for K, because the microcontroller's computing resources are limited. The smaller the value of K, the less computation is required.

[0042] Step 2: Initialize cluster centers: Randomly select K data points as initial cluster centers.

[0043] Step 3: Assign data points to clusters: For each data point, calculate its distance to each cluster center, and then assign the data point to the cluster containing the nearest cluster center.

[0044] Step 4: Update cluster centers: For each cluster, calculate the mean of all data points within the cluster and update it as the cluster center.

[0045] Step 5: Repeat steps 3 and 4: Repeatedly assign data points and update cluster centers until the stopping condition is met: the change in cluster centers is less than the set threshold or the number of iterations reaches the maximum value.

[0046] Step 6: Analyze Clusters: Analyze the size of each cluster and the data points of its members. Typically, outlier data points will form a very small cluster or be far from other clusters.

[0047] Step 7: Remove outliers: Set thresholds based on cluster size and distance criteria, and remove data that does not meet the thresholds.

[0048] Because the present invention employs the above-mentioned technical means, it has the following beneficial effects:

[0049] Compared to single-channel, multi-channel design comprehensively analyzes data from multiple sensors, which not only maintains the detection accuracy of the sensors but also expands their range and provides more detailed flow field information, resulting in higher reliability and accuracy for flow detection.

[0050] Simulation analysis and verification show that the asymmetrical distribution of upstream and downstream temperature measuring resistors can improve the threshold for the sensor to reach saturation, effectively increase the sensor's range, and determine the optimal position distribution.

[0051] The temperature sensing resistor and the heating resistor are installed by winding them around the measuring pipe, which has high heating and temperature measurement efficiency. Furthermore, the heating resistor and the temperature sensing resistor do not come into direct contact with various fluids, thus avoiding long-term exposure to fluids containing impurities or corrosive substances, which could lead to sensor wear and drift.

[0052] Three abnormal data processing methods were adopted to remove abnormal data layer by layer, which further improved the accuracy of the data.

[0053] This invention utilizes the following technical means:

[0054] I. Sensor Structural Innovation

[0055] 1. Multi-channel asymmetric design: 5 parallel circular pipes (10mm inner diameter, 180mm length, 1 component measurement in the middle and 4 flow measurement around the perimeter) are connected in parallel. The flow distribution is optimized by a fluid distribution algorithm (derived from Darcy's law) to ensure uniform flow distribution in each channel.

[0056] Asymmetric temperature measurement layout: The upstream temperature measuring resistor is 40mm away from the heating resistance wire, and the downstream temperature measuring resistor is 60mm away from the heating resistance wire (simulation verification shows that this can delay the saturation phenomenon).

[0057] 2. Integrated component detection

[0058] The central pipeline is equipped with a platinum thin film probe integrated with MEMS technology, which measures the thermal conductivity using a transient heating method to deduce the composition.

[0059] II. Circuit and Control Improvements

[0060] Improved Wheatstone bridge:

[0061] The parallel dual-branch design (R2-R5 resistor group) can still work even if one branch fails.

[0062] PID temperature control:

[0063] The temperature is calculated by detecting the resistance of the heating wire in real time, and then compared with the set value to output a PWM signal to adjust the current.

[0064] III. Level 3 Outlier Removal Algorithm

[0065] Standard deviation method: Remove data whose deviation from the mean is greater than 3σ.

[0066] Box plot method: Remove data that exceeds the range [Q1-1.5IQR, Q3+1.5IQR].

[0067] K-means clustering (K≤3): Removes outliers that are small clusters or far from the cluster center.

[0068] The following beneficial effects can be achieved:

[0069] 1. Performance Improvement

[0070] Range expansion: Multi-channel parallel connection and asymmetric layout increase saturation threshold by 200%.

[0071] Accuracy Guarantee: Three-stage data filtering ensures effective data error < ±1.5% FS

[0072] Enhanced reliability: Improved redundancy design of the bridge reduces the risk of circuit failure.

[0073] 2. Application Value

[0074] Suitable for monitoring the flow of corrosive gases (thanks to its aluminum material and insulating encapsulation) and the composition of mixed gases such as hydrogen and natural gas.

[0075] It supports dual output via Wi-Fi and serial port to meet the needs of industrial IoT. Attached Figure Description

[0076] Figure 1This is a three-dimensional schematic diagram of the structure of one embodiment of this application;

[0077] Figure 2 This is a schematic cross-sectional view of one embodiment of this application;

[0078] Figure 3 This application describes a flow signal acquisition circuit.

[0079] Figure 4 This is a schematic diagram of the installation of a signal acquisition circuit according to an embodiment of this application;

[0080] Figure 5 This is a hardware circuit flowchart of this application;

[0081] Figure 6 This is a flowchart of a main control chip according to this application;

[0082] The diagram is labeled as follows: 1-Flow inlet, 2-Flow inlet cavity, 3-Upstream temperature measuring resistor installation position, 4-Heating resistance wire installation position, 5-Downstream resistance wire installation position, 6-Flow outlet cavity, 7-Flow outlet, 8-Flow measurement pipe, 9-Component measurement pipe, 10-Probe; The flow measurement pipes include flow measurement pipe 1, flow measurement pipe 2, flow measurement pipe 3, flow measurement pipe 4, and flow measurement pipe 5. Detailed Implementation

[0083] The embodiments of the present invention will be described in detail below. Although the present invention will be described and illustrated in conjunction with some specific embodiments, it should be noted that the present invention is not limited to these embodiments. On the contrary, any modifications or equivalent substitutions made to the present invention should be covered within the scope of the claims of the present invention.

[0084] Furthermore, to better illustrate the present invention, numerous specific details are set forth in the following detailed embodiments. Those skilled in the art will understand that the present invention can be practiced without these specific details.

[0085] In this example, the multi-channel thermal flow sensor mainly includes: a pipeline for measuring overall flow and components, hardware circuitry, a microcontroller system, a mobile terminal, and a PC terminal.

[0086] In this example, fluid flows into the flow inlet chamber 2 from an external pipe, and then branches into the flow measurement pipe and the component measurement pipe. The four flow measurement pipes 8 and one component measurement pipe 9 are the channels through which the fluid flows, and their cross-sections are circular. The flow rate distribution in each measurement pipe is determined by the structure of the flow inlet chamber and the relative position of the flow inlets. When the fluid passes through the flow measurement pipe, it alters the original temperature field generated by the heating resistance wire. The upstream and downstream temperature measurement resistance wires detect the temperature difference, and based on the relationship between the temperature difference and the flow rate, the flow rate data is obtained.

[0087] To evaluate the distribution of mass flow rate in a measured pipeline, this example uses the relative standard deviation method to describe the dispersion of mass flow rate.

[0088] Define the mass flow rate deviation coefficient: Where, m i Let i be the mass flow rate of channel i. This represents the average quality flow rate across all channels.

[0089] The mass flow rate distribution deviation is defined as: Where n is the number of channels. Uneven flow distribution leads to uneven energy utilization across channels. Darcy states that for parallel circuits, the potential difference across the circuit terminals is uniform. Similarly, in parallel fluid channels, the friction loss across each channel should also be consistent. The formula is expressed as: λ is the friction loss coefficient, l is the length of the measuring channel, d is the inner diameter of the channel, V is the fluid velocity, and g is the acceleration due to gravity. Based on the principle that the friction losses in the main channel and the sub-channels are equal, we can derive:

[0090]

[0091] Where λ1 and λ2 are the friction coefficients of the main channel and the branch pipe, respectively; L and l are the lengths of the main channel and the branch pipe, respectively; D and d are the characteristic dimensions of the main channel and the branch pipe, respectively; ξ′ and σξ″ are the sum of the friction coefficients generated by the branch pipe extending into the main channel and the local friction coefficients of the branch channel, respectively; V1 and V2 are the average velocity of the fluid in the main channel and the average velocity of the fluid in the branch channel, respectively; and g is the acceleration due to gravity.

[0092] Fluid flow theory and experimental studies have shown that, given a fixed design and material for the main channel, if the construction, dimensions, components, and connection methods of the branch channels to the main channel remain unchanged, the sum of the local drag coefficient and the total drag coefficient ξ′, σξ″, can be considered constant values. Therefore:

[0093] q1 = cq2

[0094] Where q1 represents the inlet flow rate of the main channel, q2 represents the total flow rate of the branch measuring pipes, and c is a constant, considered as a proportionality coefficient. Therefore, by measuring and obtaining the coefficient, the relationship between the flow rate through the inlet channel and the branch inlet pipe can be obtained. Then, by changing the relative positions of each channel and their corresponding cross-sectional area parameters, and through simulation verification and comparison, the β value of each pipe can be obtained. i The result with the smallest value is used to determine the pipeline model and parameters in this example, such as... Figure 1 As shown, the flow rate is distributed almost evenly among the branch pipes. If a large error occurs in one pipe compared to the others, it can be detected and corrected in a timely manner, thus achieving high precision and extended range.

[0095] Furthermore, the flow sensor structure is entirely made of aluminum, and the parallel circular pipe has an outer diameter of 10.5 mm, an inner diameter of 10 mm, and a length of 180 mm. Aluminum has excellent corrosion resistance, protecting against corrosion from common gases and ensuring the long-term stability and accuracy of the sensor. On the other hand, aluminum is easy to process, reducing manufacturing costs.

[0096] Furthermore, a platinum thin film probe integrated with MEMS technology is provided in the center of the component measurement pipe. The probe is formed into a micron-scale sensitive structure through photolithography and etching technology, and its arrangement direction is parallel to the gas flow direction.

[0097] The above technical solution reduces the disturbance of the probe to the gas flow and improves the measurement accuracy at higher gas flow rates.

[0098] Furthermore, the probe is connected to a relay and controlled by a microcontroller to achieve transient heating by applying a constant power pulse of 0.5–2 seconds. Subsequently, the resistance change of the probe is detected at a certain frequency to obtain the temperature change curve of the probe.

[0099] The above technical solution allows for the measurement of the thermal conductivity of a gas, and further, the determination of its composition. The principle is as follows: In an infinitely large, static gas medium, the probe temperature rise ΔT satisfies the following relationship with time t:

[0100]

[0101] Where C is a constant, q is the pulse power, and k is the gas thermal conductivity. It can be seen that the slope of the temperature rise curve, d(ΔT) / d(Int), is inversely proportional to the thermal conductivity k. Essentially, the higher the gas thermal conductivity, the slower the probe temperature rises under the same heating power. Furthermore, the composition of the gas can be determined based on its thermal conductivity. The relationship between thermal conductivity and composition is as follows:

[0102] λ mix =∑(Xi·λ i )

[0103] Where, λmix Let λ be the thermal conductivity of the gas mixture, Xi be the proportion of the i-th gas, and λ be the thermal conductivity of the mixture. i Let be the thermal conductivity of the i-th gas. If there are only two gases in the mixture, then ∑(Xi) = 100%. Therefore, by solving the two equations simultaneously, the composition of the gas mixture can be obtained from its thermal conductivity. It is evident that this method is only applicable to the measurement of gas composition in a mixture of two single gases. Since there is a relationship between probe resistance change and thermal conductivity, and a relationship between thermal conductivity and gas composition, we can deduce that there is also a relationship between probe resistance change and gas composition. This relationship can be fitted experimentally.

[0104] Each flow measurement pipe has a heating resistance wire wound around its center, and temperature sensing resistance wires wound asymmetrically on both sides. The distance between the heating resistance wire and the temperature sensing resistance wires directly affects the accuracy of the flow sensor. During the design process, sufficient distance must be maintained between the heating and temperature sensing resistance wires to allow the fluid ample time and space for heat exchange with the surrounding environment after flowing over the heating resistance wire. However, this distance cannot be too large, as it would increase the system's response time. Simulation analysis at different distances determined that the upstream and downstream temperature sensing resistance wires are asymmetrically distributed relative to the heating resistance wire. The distance between the upstream and downstream resistance wires and the heating resistance wire is 40mm, and the distance is 60mm. The distance between the upstream and downstream resistance wires and both ends of the parallel circular pipe is 30mm. The slope of the function graph of the temperature difference of the temperature sensing element versus flow velocity is the largest at this point, indicating that the sensor exhibits high sensitivity, high response efficiency, and a large measurement range.

[0105] Furthermore, the heating resistor is a nickel-chromium resistance wire with a diameter of 0.1 mm, and the temperature measuring resistor is a platinum resistance wire with a diameter of 0.05 mm. The surfaces of both the heating resistor wire and the temperature measuring resistor wire are insulated.

[0106] The above technical solution uses nickel-chromium resistance wire as the heating resistor, which has a high temperature coefficient of resistance and good high temperature resistance, and a fast heating speed. Platinum resistance wire is used as the temperature measuring resistor, which has good stability, very linear temperature characteristics, and high temperature measurement accuracy.

[0107] The heating resistance wire and the temperature measuring resistance wire of the flow measurement pipeline are both installed at locations where the outer diameter of a small section of the pipeline is slightly smaller than that of other locations, and thermally conductive material is placed between the heating resistance wire and the temperature measuring resistance wire and the circular pipeline.

[0108] The above technical solution offers two advantages: firstly, it facilitates the installation and positioning of the temperature sensing resistance wire and the heating resistance wire; secondly...

[0109] To avoid friction between the heating resistance wire, temperature sensing resistance wire and the measuring pipe, which could damage the insulation layer, and to increase the heat conduction efficiency at the installation location of the heating resistance wire and temperature sensing resistance wire.

[0110] Furthermore, during the sensor's measurement process, the flow measurement pipe is completely covered with insulating cotton to prevent external airflow from affecting the measurement accuracy.

[0111] Furthermore, the connection between the flow inlet cavity and the flow measurement pipe, as well as the connection between the flow outlet cavity and the flow measurement pipe and the component measurement pipe, are all provided with rounded corners, such as... Figure 2 As shown. The rounded corners facilitate fluid inflow and outflow from the measuring pipe, further ensuring the uniformity of flow rate in each channel.

[0112] Furthermore, the heating resistor is connected to the power supply via a relay, and the relay is controlled by a PWM signal output from a microcontroller.

[0113] Furthermore, the temperature difference signal of the measuring resistors at both ends of the measuring tube is acquired by a signal acquisition circuit. Since the acquired temperature difference signal is very weak, usually at the millivolt level, it needs to be amplified by a differential signal amplifier circuit. However, noise will inevitably be generated during the amplification process, affecting data processing. Therefore, the noise needs to be filtered out by a filter circuit. Finally, the analog signal is converted into a digital signal by the microcontroller's A / D converter and processed inside the microcontroller.

[0114] Furthermore, the signal acquisition circuit employs an improved Wheatstone bridge. Since traditional Wheatstone bridges still have certain shortcomings in terms of accuracy and prevention of circuit failure, the above solution proposes a signal acquisition circuit with protective features.

[0115] It consists of platinum resistance wire R S1 R S2 Together with precision resistors R2, R3, R4, and R5, it forms the circuit. The voltage signal acquired by the improved circuit is as follows:

[0116]

[0117] When R S1 =R S2 =R2=R3=R4=R5=R0, and ΔR S1 =ΔR, ΔR S2 When =-ΔR, the above formula can be simplified to:

[0118]

[0119] Through the above technical solution, the parallel circuit composed of R2, R3, R4 and R5 can control the circuit resistance value relatively easily. Furthermore, if one resistor in a parallel circuit fails, the other resistor can still play a role, preventing the circuit from completely failing and thus protecting the circuit.

[0120] Furthermore, in the above measurement scheme, based on the temperature of the fluid being measured, the temperature of the heating resistor is generally set to be higher than the fluid temperature by a certain value to ensure the measurement accuracy of the sensor. To maintain a constant temperature of the central heating resistor wire and prevent the fluid flow from carrying away a large amount of heat, which would affect the measurement accuracy, the current through the heating resistor wire is measured by the microcontroller's A / D converter, and the resistance value of the heating resistor is calculated. Based on the certain correlation between resistance value and temperature, the current temperature of the resistor is finally obtained. Then, the current resistor temperature and the desired temperature are used for PID calculation, and a PWM signal is output to control the operation of the relay, thereby ensuring that the temperature of the central heating resistor remains constant.

[0121] Furthermore, after processing and analysis, the data measured by the flow sensor is displayed on the terminal to facilitate real-time monitoring of the pipeline's flow and component data. In this example, the flow and component data, after anomaly data removal, are transmitted to the mobile terminal and PC via a WIFI module and serial communication, respectively. This enables real-time observation of measurement data both on-site and remotely, meeting the needs of some special flow measurement scenarios.

[0122] Furthermore, in this example, the traffic data detected by the hardware circuit will inevitably contain some abnormal data that deviates significantly from the actual value. These abnormal data need to be eliminated to the greatest extent possible through the algorithm process of the microcontroller system.

[0123] The processing flow consists of the following three steps:

[0124] (I) Preliminary data processing using the standard deviation method. In the measurement experiment, the test data follows a normal distribution. Approximately 68% of the data falls within one standard deviation of the mean, about 95% falls within two standard deviations, and 99.7% falls within three standard deviations. First, the mean (μ) and standard deviation (σ) of the collected data sets are calculated. Then, the threshold for outliers is set at three standard deviations. If the data falls outside three standard deviations, the corresponding data is removed, and the remaining data are retained.

[0125] Standard deviation is usually unknown, but experimental standard deviation can be used instead. In a set of n independent and repeated observations, the k-th observation X... k The difference between the digital quantity converted from the analog-to-digital converter (ADC) by the microcontroller and the arithmetic mean of the test data is called the residual error V. k ,but:

[0126]

[0127] Therefore, the experimental standard deviation s(X) k )for:

[0128]

[0129] in This represents the arithmetic mean of a set of n observations.

[0130] (II) Further processing using box plots. Arranging the data from the first step in ascending order reveals the first quartile (Q1): the number at the 25th percentile; the median: the number at the 50th percentile; and the third quartile (Q3): the number at the 75th percentile. The difference between the third quartile and the first quartile is called the interquartile range (1QR), an important statistic for measuring the dispersion of the data. That is:

[0131] IQR = Q3 - Q1

[0132] The upper edge of the data is: M s =Q1 + 1.5 × IQR

[0133] The bottom edge of the data is: M x =Q3-1.5×IQR

[0134] Where Q1 represents the number in the 25th percentile after the data is sorted in ascending order, Q3 represents the number in the 75th percentile, and IQR represents the difference between Q3 and Q1.

[0135] Data falling outside the upper and lower edges of the data set is considered outlier, and the corresponding data is removed, while the remaining data is retained.

[0136] (III) The final step of data processing is to use k-means clustering.

[0137] Step 1: Determine the number of clusters: Choose a relatively small value for K, because the microcontroller's computing resources are limited. The smaller the value of K, the less computation is required.

[0138] Step 2: Initialize cluster centers: Randomly select K data points as initial cluster centers.

[0139] Step 3: Assign data points to clusters: For each data point, calculate its distance to each cluster center, and then assign the data point to the cluster containing the nearest cluster center.

[0140] Step 4: Update cluster centers: For each cluster, calculate the mean of all data points within the cluster and update it as the cluster center.

[0141] Step 5: Repeat steps 3 and 4: Repeatedly assign data points and update cluster centers until the stopping condition is met: the change in cluster centers is less than the set threshold or the number of iterations reaches the maximum value.

[0142] Step 6: Analyze Clusters: Analyze the size of each cluster and the data points of its members. Typically, outlier data points will form a very small cluster or be far from other clusters.

[0143] Step 7: Remove outliers: Set thresholds based on cluster size and distance criteria, and remove data that does not meet the thresholds.

[0144] Data that meets all three error criteria is considered valid and saved. Data that fails to meet any one of the criteria should be discarded or saved separately. To avoid the possibility of randomness affecting accuracy, multiple measurements should be performed. Finally, outlier data is removed, and the average of the remaining valid data is taken as the result of a single measurement.

[0145] It is important to note that the number of observations in each group should be determined based on the sampling frequency, the specific requirements of the system, and the computing power of the microcontroller. It should not be too many or too few. If the sampling frequency is too high, it will affect the working life and sensitivity of the sensor. If the gas flow rate being measured changes greatly, if the sampling frequency is too low, it will affect the accuracy of the statistical results of the extracted data and the response time. An appropriate number of measurements should be selected according to the specific circumstances.

[0146] It will be understood by those skilled in the art that the above descriptions are merely preferred examples of the invention and are not intended to limit the invention. Although the invention has been described in detail with reference to the foregoing examples, those skilled in the art can still modify the technical solutions described in the foregoing examples or make equivalent substitutions for some of the technical features. All modifications and equivalent substitutions made within the spirit and principles of the invention should be included within the scope of protection of the invention.

Claims

1. An asymmetric multi-channel thermal flow sensor integrating component detection, characterized in that, include: The flow inlet (2) and flow outlet (6) are respectively provided with a flow inlet (1) and a flow outlet (7); The flow measurement pipe group and component measurement pipe (9) connect the flow inlet and flow outlet of the oral cavity; The component measurement pipe is located in the middle of all the pipes, while the flow measurement pipe is located around the perimeter. A probe (10) is installed in the middle of the component measurement pipe; Each flow measurement circular pipe is equipped with a heating resistance wire, and temperature measuring resistance wires are respectively installed on the upstream and downstream sides of the heating resistance wire. The first distance between the upstream temperature measuring resistance wire and the heating resistance wire is less than the second distance between the downstream temperature measuring resistance wire and the heating resistance wire.

2. The sensor as described in claim 1, characterized in that, The flow measurement pipe assembly comprises four parallel circular pipes (8), with an inner diameter of 10 mm and a length of 180 mm, a first distance of 40 mm, and a second distance of 60 mm.

3. The sensor as described in claim 1, characterized in that, The component measurement pipe (9) has an inner diameter of 10 mm and a length of 180 mm, with the probe located in the center of the pipe and the probe being arranged in a direction parallel to the gas flow direction.

4. The sensor as described in claim 1, characterized in that, The connection between the flow inlet cavity (2) and the flow measurement pipe assembly, as well as the connection between the flow outlet cavity (6) and the flow measurement pipe assembly, are all provided with rounded corner transition structures.

5. The sensor as described in claim 1, characterized in that, The heating resistance wire (4) is made of nickel-chromium alloy, the temperature measuring resistance wire is made of platinum, and the component measurement probe is a platinum thin film probe integrated by MEMS process, and their surfaces are all covered with an insulating layer.

6. The sensor as described in claim 1, characterized in that, The heating resistance wire and the temperature measuring resistance wire are wound around the outer wall of the pipe, and the space between them and the pipe is filled with heat-conducting material, and the pipe is covered with an insulation layer.

7. The sensor as described in claim 1, characterized in that, It also includes an improved Wheatstone bridge circuit, which comprises a first branch and a second branch connected in parallel: First branch: Resistors R2 and R3 are connected in series and then connected to the platinum resistance wire R. S1 Series; Second branch: Resistors R4 and R5 are connected in series and then connected to the platinum resistance wire R. S2 Series; When R S1 =R S2 =R2=R3=R4=R5=R0, and ΔR S1 =ΔR,ΔR S2 When =-ΔR, the above formula simplifies to: Where ΔR S1 For resistor R S1 Change value, ΔR S2 For resistor R S2 The change value, ΔR represents the change in the measured resistance, U is the input voltage, and ΔU is the voltage collected by the circuit.

8. The sensor according to claim 1, characterized in that: The control module includes a PID controller, which calculates the temperature by detecting the resistance of the heating wire in real time, and outputs a PWM signal after comparing it with the set temperature. The PWM signal regulates the heating current via a relay to maintain a constant temperature for the heating resistance wire.

9. A signal processing method for the sensor according to any one of claims 1-6, characterized in that, include: The temperature difference signals from each channel are collected, amplified, filtered, and converted by an A / D converter to obtain digital signals. Three-layer outlier removal for digital signals: (a) Standard deviation method: Data that deviates from the mean by more than 3 standard deviations are removed; (b) Box plot method: Calculate the interquartile range (IQR) of the data (Q3 - Q1), and remove data that are outside the range of [Q1 + 1.5 × IQR, Q3 - 1.5 × IQR]. (c) K-means clustering: Set K≤3, and after clustering the data, remove data with fewer than the threshold or far from the cluster center; Finally, the average value of the valid data is taken to calculate the pipeline flow rate and gas composition, and the results are output via WIFI / serial port.

Images