A dynamic hysteresis compensation method and system for a micro-flow sensor

Abstract

The application discloses a dynamic hysteresis compensation method and system of a micro-flow sensor, and belongs to the field of micro-flow control. The application removes interference by identifying the overall model of the flow sensor and combining theoretical analysis; constructs an ideal model of the flow sensor and collects actual output data; identifies a dynamic hysteresis compensation unit through a genetic algorithm, and performs hysteresis compensation on the flow sensor; and switches reasonably between the actual flow sensor output flow and the hysteresis compensation output through mutual correction algorithms, which can improve the response performance of the flow output and ensure accurate flow output. The application is suitable for micro-flow control in the field of micro-flow control, micro-thruster system flow control in the field of aviation and aerospace, and micro-flow control in the field of semiconductors. The application solves the problem of large hysteresis of the micro-flow sensor, compensates the dynamic hysteresis of the flow response, improves the response performance of the micro-flow sensor, and improves work efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of flow control valve and micro flow sensor system hysteresis flow compensation, in particular to a dynamic hysteresis compensation method and system of micro flow sensor, belongs to the field of micro flow control. BACKGROUND

[0002] Pneumatic technology uses compressed gas as working medium to transfer energy and signal and control, which is an important technology to realize production automation. Since the micro gas flow control technology was first proposed in the early 1990s, only after a few decades of development, it has made great progress and remarkable results, and is the frontier and hotspot in the current science and technology field. Micro flow control technology can be applied to micro thruster systems in the field of aerospace and semiconductor field. Piezoelectric actuator can realize high-precision control of fluid because of its small power consumption, fast response, high precision and other advantages, and piezoelectric proportional valve composed of valve, which has been more and more widely used in micro fluidic field in recent years.

[0003] Responsiveness has always been the target pursued in the field of micro fluidics, but the accuracy and real-time performance of gas flow measurement depend on the performance of the flow sensor. The micro flow measurement sensor on the market generally has a large inertia hysteresis phenomenon. This is due to the mechanism of dynamic measurement of micro flow sensor. The laminar flow differential pressure type flow sensor needs a certain length of laminar element because of the measurement principle, and the flow flow establishment time is long. The thermal flow sensor has a transition process because of the reestablishment of thermal balance, and has a measurement hysteresis. In order to obtain more accurate and real-time flow signal, the dynamic hysteresis of the flow sensor needs to be compensated. The compensation method of the previous sensor model establishment and identification is not suitable for micro flow sensor. The mechanism of micro flow measurement makes it difficult to model dynamically, and the black box identification method is difficult to separate because of the coupling relationship with the execution valve, resulting in inaccurate identification. Therefore, a new solution to the dynamic hysteresis problem of micro flow sensor is needed, and the hysteresis is compensated in combination with the system. SUMMARY

[0004] In view of the problems of large hysteresis and slow response of micro flow sensor system, the main purpose of the present application is to disclose a dynamic hysteresis compensation method and system of micro flow sensor. Through the dynamic hysteresis compensation unit and the mutual correction algorithm, the dynamic hysteresis error of micro flow sensor flow response is compensated, and the response performance of micro flow sensor is improved without losing the accuracy of micro flow sensor.

[0005] The purpose of the present application is realized by the following technical scheme:

[0006] The application discloses a dynamic hysteresis compensation method and system of a micro-flow sensor.

[0007] The dynamic hysteresis compensation system of the micro-flow sensor comprises a gas source, a flow sensor, a microprocessor and a flow control valve.

[0008] The flow control valve is an execution unit of flow regulation.

[0009] The flow sensor is used for collecting actual flow data output by the flow control valve.

[0010] The microprocessor comprises a dynamic hysteresis compensation unit, a mutual correction algorithm and microprocessor calculation.

[0011] The application discloses a dynamic hysteresis compensation method of a micro-flow sensor.

[0012] Step 1: a sine wave form of flow is generated through the flow control valve, and the flow is measured through the flow sensor to obtain input and output data, and a whole model containing the flow control valve and the flow sensor is identified through a gradient correction method.

[0013]

[0014] In the formula, A1…A i , B1…B j are model parameters.

[0015] Step 2, remove the predictable lag part of the flow control valve model and the flow sensor from the identified overall model according to the theoretical model analysis of the flow control valve and the flow sensor, to obtain a desired ideal flow sensor model satisfying the given performance requirements, and the discrete model is as follows:

[0016]

[0017] In the formula, a1, b1, b2 are model parameters.

[0018] Step 3, construct multiple sets of input signals composed of lift flow change step signals and return flow change step signals. The multiple sets of input signals are respectively applied to the sensor system and the desired ideal model established in step 2, to obtain actual flow response signals Q(k) and expected standard response signals Qc(k) respectively; the actual flow response signals Q(k) will obtain compensated signals Qb(k) through the established dynamic lag compensation unit Gc(z -1 ); the fitness function F(k) of the genetic algorithm takes the absolute value of the difference between the compensated signals Qb(k) and the standard response signals Qc(k), and then the genetic algorithm is used to continuously optimize the parameters in the dynamic lag compensation unit Gc(z -1 ) in the direction of the minimum fitness function, so that the compensated signals Qb(k) continuously approach the expected standard response signals Qc(k); -1

[0019]

[0020] F(k) = |Qc(k) - Qb(k)| (4)

[0021] In the formula, a c1 , a c2 , b c1 , b c2 are model parameters of the dynamic lag compensation unit Gc(z -1 ); F(k) is the fitness function.

[0022] Step 4, the initial value of the model parameters of the dynamic lag compensation unit Gc(z -1 ) established in step 3 will be determined by the initialization population step of the genetic algorithm, and the upper and lower limits [a cmin , a cmax ], [b cmin , b cmax ] of the Gc(z -1 ) parameters will be determined according to the parameters of the identified overall model. And set the population size g, crossover probability p c , mutation probability p b ​The fitness function is shown as equation (3), and the three termination conditions are maximum genetic generation Dmax, the best chromosome unchanged for n times, and the variance of population fitness value < preset variance qj.

[0023] Step 5, the verification data set is established, containing the complement of the sinusoidal signal flow, the actual flow sensor output Q(k) is given to the dynamic lag compensation unit to obtain the compensated signal Qb(k), and compared with the given sensor dynamic performance response requirement, if it meets the requirement, no further optimization is needed, that is, the final dynamic lag compensation unit is obtained, if it does not meet the requirement, the identification data set is expanded, and the dynamic compensation unit is revised again until it meets the given response performance requirement.

[0024] Step 6, through the mutual correction algorithm, reasonable switching is carried out between the actual flow sensor output flow and the lag compensation output, and the response performance of the micro flow sensor is improved.

[0025] The specific steps of the mutual correction algorithm are as follows:

[0026] Step 6.1, whether the lag compensation flow Qb and the expected set flow Qset meet the following formula (5) is compared, if it meets, step 6.2 is carried out, if it does not meet, the mutual correction algorithm output correction flow Qy is equal to the lag compensation flow Qb;

[0027] Qb≥k1·Qset (5)

[0028] Qy=Qb(Qb<k1·Qset) (6)

[0029] In the formula, k1 is the response judgment coefficient, and the value range is 63%, 95%, and 99%.

[0030] Step 6.2, the lag compensation flow Qb and the actual flow response Q are further compared, if the absolute value of the difference between the two is less than the preset threshold value ε, step 6.4 is entered, if it does not meet, the mutual correction algorithm output correction flow Qy is equal to the lag compensation flow Qb;

[0031] Qy=Qb(|Qb-Q|>ε) (7)

[0032] Step 6.3, the threshold value in step 6.2 is linearly fitted by the maximum prediction error of the lag compensation flow Qb and the actual flow response Q at steady state under different step responses, and it is found through the empirical formula (8) that the error is usually less than 3% of the actual flow response Q at steady state, so the threshold value ε can be approximately replaced by this value, which meets the response of the lag compensation, and can switch when the two flows are close, and the accuracy of the flow sensor output is ensured.

[0033] e max =k2·U (8)

[0034] ε=k3·Q (9)

[0035] In the formula, e max To represent the maximum prediction error, U is the input signal value of the flow control valve, k2 is the fitting coefficient, and k3 is the valve value correlation coefficient, with values ​​ranging from 5%, 3%, and 1%.

[0036] Step 6.4: When the conditions in steps 6.1 and 6.2 are met, the mutual calibration algorithm outputs a calibrated flow rate Qy equal to the actual flow response Q. During switching, continuously judging the conditions in step 2 would impact the algorithm's computational power and cause noise-induced fluctuations in the flow sensor output. Performing only one judgment would lead to misjudgments due to sensor noise. Therefore, the method of multiple judgments and switching or delayed de-jitter judgment as described in step 6.5 is adopted.

[0037] Qy=Q(Qb≥k1·Qset and |Qb-Q|<ε) (10)

[0038] Step 6.5, Multiple Judgment Switching: If the absolute value of the difference between the calculated hysteresis compensation flow Qb and the actual flow response Q within 5 consecutive discrete periods T is within the threshold ε, then switching is performed, and no further judgment is made. This method offers better reliability and avoids outlier issues to some extent. Delayed Debounce Judgment: After a judgment reaches the threshold ε, a delay t is made to check whether the absolute value of the difference in the next 3 discrete periods is within the threshold ε. If so, switching is performed. This algorithm consumes minimal computational power and eliminates the jitter caused by switching.

[0039] Beneficial effects:

[0040] 1. The present invention discloses a dynamic hysteresis compensation method and system for a micro-flow sensor, which solves the large hysteresis problem of the micro-flow sensor by establishing a dynamic hysteresis compensation unit, compensating for the dynamic hysteresis of its flow response, and improving its responsiveness.

[0041] 2. The present invention discloses a dynamic hysteresis compensation method and system for a micro-flow sensor, which reasonably switches between the actual flow sensor output flow and the hysteresis compensation output through a mutual calibration algorithm, thereby ensuring both the responsiveness of the micro-flow sensor output and the accuracy of the output flow. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the structure of a dynamic hysteresis compensation system for a micro-flow sensor disclosed in this invention;

[0043] Figure 2 This is a flowchart of a dynamic hysteresis compensation method for a micro-flow sensor disclosed in this invention;

[0044] Figure 3 The schematic diagram of the identification method of the dynamic lag compensation unit;

[0045] Figure 4 The flowchart of the mutual correction algorithm;

[0046] Figure 5 The schematic diagram of the fitness function value changing with the genetic generation in the embodiment;

[0047] Figure 6 The schematic diagram of the flow sensor lag compensation flow and the actual flow in the embodiment;

[0048] Figure (a) is the overall comparison diagram of the flow sensor lag compensation flow and the actual flow; Figure (b) is the enlarged comparison diagram of the flow sensor lag compensation flow and the actual flow in the lift section; Figure (c) is the enlarged comparison diagram of the flow sensor lag compensation flow and the actual flow in the return section;

[0049] Figure 7 The threshold fitting schematic diagram in the embodiment;

[0050] Figure 8 The comparison diagram of the mutual correction output correction flow and the actual flow in the embodiment. DETAILED DESCRIPTION

[0051] The present application will be described in detail below with reference to the accompanying drawings and embodiments. The technical problems solved by the technical solutions of the present application and the beneficial effects are also described. It should be pointed out that the described embodiments are only intended to facilitate the understanding of the present application and do not limit the present application in any way.

[0052] Since the micro-thrust cold gas thruster has strict requirements on the responsiveness of the thrust, and the thrust is measured by the flow sensor to calculate the flow, the response performance of the flow sensor will directly affect the quality of the micro-thrust cold gas thruster. The present embodiment discloses a dynamic lag compensation method and system for a micro-flow sensor, which improves the responsiveness of the micro-flow sensor in the micro-thrust cold gas thruster through dynamic lag compensation. The sensor range is 0-5sccm.

[0053] As shown in Figure 1 A dynamic lag compensation method and system for a micro-flow sensor, characterized in that the system comprises: a gas source, a flow sensor, a microprocessor, and a flow control valve; the gas source outlet is connected to the flow control valve, the flow control valve outlet is connected to the flow sensor, the flow sensor signal output interface is connected to the input interface of the microprocessor, and the output interface of the microprocessor is connected to the input interface of the flow control valve, for adjusting the opening size of the flow control valve.

[0054] The flow control valve is a flow regulating execution unit; the flow sensor is used to collect the actual flow data output by the flow control valve; the microprocessor includes a dynamic hysteresis compensation unit algorithm, a mutual correction algorithm, and microprocessor calculation; the dynamic hysteresis compensation unit is used to compensate the hysteresis in the actual flow signal output by the flow sensor, obtaining a hysteresis compensation flow. The mutual correction algorithm is used to reasonably switch between the actual flow sensor output flow and the hysteresis compensation output, ensuring both responsiveness and accuracy; the microprocessor processes and calculates according to the dynamic hysteresis compensation unit and the mutual correction algorithm, and sets the size of the flow regulating flow control valve output flow.

[0055] As shown in Figure 2 A dynamic hysteresis compensation method for a micro flow sensor, based on a dynamic hysteresis compensation system for a micro flow sensor, includes the following steps:

[0056] Step 1, generate a sinusoidal wave form (such as Q = 2sin(0.5πt-0.5π) + 2) flow through the flow control valve, and measure it through the flow sensor to obtain input-output data, and identify the overall model containing the execution structure and the flow sensor through the gradient correction method;

[0057]

[0058] Step 2, remove the flow control valve model and the predictable hysteresis part of the flow sensor from the identified overall model by combining the theoretical model analysis of the execution structure and the flow sensor, obtaining the desired ideal flow sensor model that meets the given performance requirements, and the discrete model is as follows:

[0059]

[0060] In the formula, a1 = 0.046, b1 = -0.014, and b2 = -0.954 are model parameters.

[0061] Step 3, as shown in Figure 3As shown, the multiple sets of input signals consisting of lift flow variation step signals and backflow variation step signals are constructed. The multiple sets of input signals are respectively applied to the sensor system and the desired ideal model established in step 2 to obtain actual flow response signals Q(k) and expected standard response signals Qc(k) respectively; the actual flow response signals Q(k) will pass through the established dynamic lag compensation unit Gc(z-1) to obtain compensated signals Qb(k); the fitness function F(k) of the genetic algorithm takes the absolute value of the difference between the compensated signals Qb(k) and the standard response signals Qc(k), and then the genetic algorithm is used to continuously optimize the parameters in the dynamic lag compensation unit Gc(z-1) in the direction of the minimum fitness function, so that the dynamic lag compensation unit Gc(z-1) makes the compensated signals Qb(k) continuously close to the expected standard response signals Qc(k);

[0062]

[0063] F(k) = |Qc(k) - Qb(k)| (14)

[0064] wherein a c1 = 0.167, a c2 = -0.164, b c1 = -1.889, and b c2 = 0.892 are model parameters of the dynamic lag compensation unit Gc(z -1 ).

[0065] In step 4, the initial values of the model parameters of the dynamic lag compensation unit Gc(z -1 ) established in step 3 are determined by the initialization population step of the genetic algorithm, and the upper and lower limits [a cmin = -5, a cmax = 5], [b cmin = -5, b cmax = 5] of the Gc(z -1 ) parameters are determined according to the parameters of the overall model. The population number g = 50, the crossover probability p c = 0.8, the mutation probability p b = 0.01, the fitness function as shown in equation (14), and the result of the genetic generation transformation are shown in Figure 5 The three termination conditions are set as the maximum genetic generation Dmax = 10000, the best chromosome unchanged for 1000 times continuously, and the variance of the population fitness value <0.0025, respectively.

[0066] Step 5, the verification data set is established, such as the sine wave signal flow (Q = 2sin (0.5pi t-0.5pi) + 2.5), the actual flow sensor output Q (k) is given to the dynamic lag compensation unit to obtain the compensated signal Qb (k), and the given sensor dynamic performance response requirement is compared, such as meeting the requirement, no further optimization is needed, that is, the final dynamic lag compensation unit is obtained. If it does not meet the requirement, the identification data set is expanded, and the dynamic compensation unit is revised again until it meets the given response performance requirement, and the final lag compensation result is shown in Figure 6 .

[0067] Step 6, as shown in Figure 4 , the mutual correction algorithm is used to reasonably switch between the actual flow sensor output flow and the lag compensation output.

[0068] Step 6.1, by comparing whether the lag compensation flow Qb and the expected set flow Qset meet the following formula (15), if they meet, step 6.2 is performed; if they do not meet, the mutual correction algorithm outputs the corrected flow Qy equal to the lag compensation flow Qb;

[0069] Qb≥k1·Qset (15)

[0070] Qy=Qb(Qb<k1·Qset) (16)

[0071] In the formula, k1 is the response judgment coefficient, which is 95%.

[0072] Step 6.2, further compare the lag compensation flow Qb and the actual flow response Q, if the absolute value of the difference between the two is less than the preset threshold value epsilon, step 6.4 is entered; if it does not meet, the mutual correction algorithm outputs the corrected flow Qy equal to the lag compensation flow Qb;

[0073] Qy=Qb(|Qb-Q|>ε) (17)

[0074] Step 6.3, the threshold value in step 6.2 is linearly fitted by the maximum prediction error of the lag compensation flow Qb and the actual flow response Q at steady state under different step responses, as shown in Figure 7 , it is found through the empirical formula (18) that the error is usually less than 3% of the actual flow response Q at steady state, so the threshold value epsilon can be approximately replaced by this value, which meets the response of the lag compensation and switches when the two flows are close, ensuring the accuracy of the flow sensor output.

[0075] e max =k2·U (18)

[0076] ε=k3·Q (19)

[0077] In the formula, emax For the maximum prediction error, U is the input signal value of the flow control valve, k2=0.0067 is a fitting coefficient, and k3 is a valve-related coefficient, which is 3%.

[0078] In step 6.4, when the conditions in steps 6.1 and 6.2 are met, the mutual correction algorithm outputs a corrected flow Qy equal to the actual flow response Q. When switching is performed, if the condition in step 6.2 is continuously judged, the algorithmic load may be affected, and the flow sensor output may be dithered due to noise. If the judgment is performed only once, the sensor noise may cause a false judgment. Therefore, the multiple judgment switching in step 6.5 or the delay dithering judgment method can be used.

[0079] Qy=Q(Qb≥k1·Qset and |Qb-Q|<ε) (20)

[0080] In step 6.5, multiple judgment switching: if the absolute value of the difference between the hysteresis compensation flow Qb and the actual flow response Q calculated in 5 consecutive discrete periods T is within the threshold ε, switching is performed, and no further judgment is performed. The reliability is better, and the problem of abnormal values is avoided to some extent. Delay dithering judgment: after one judgment enters the threshold ε, delay for a period of time t to judge again whether the absolute value of the difference in the next 3 discrete periods is within the threshold ε. If so, switching is performed. The algorithmic load is minimized, and the dithering caused by switching can be eliminated, as shown in Figure 8 .

[0081] The dynamic hysteresis compensation method and system for a micro-flow sensor proposed in the application solve the problem of large hysteresis of the micro-flow sensor, compensate for the dynamic hysteresis of the flow response, and improve the responsiveness. The method and system can be used in micro-flow measurement, further applied in the flow control of a micro-thruster system in the field of aerospace, and micro-fluidic technology in the field of semiconductors, to improve the response speed and further improve the work efficiency.

[0082] The specific description described above further details the purpose, technical solutions, and beneficial effects of the application. It should be understood that the above description is only a specific embodiment of the application and does not limit the protection scope of the application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the application should be included in the protection scope of the application.

Claims

1. A method of dynamic hysteresis compensation for a microflow sensor, characterized by: The method comprises the following steps, Step 1, generating a sine wave form of flow through a flow control valve, and measuring the flow through a flow sensor to obtain input and output data, and identifying an overall model containing the flow control valve and the flow sensor through a gradient correction method; wherein A1...A i , B1...B j are model parameters; Step 2, removing a flow control valve model and a predictable lag part of the flow sensor from the identified overall model by combining theoretical model analysis of the flow control valve and the flow sensor to obtain an expected ideal flow sensor model satisfying a given performance requirement, and a discrete model is as follows: In the formula, a1, b1 and b2 are model parameters; Step 3, construct a plurality of sets of input signals composed of lift flow variation step signals and backhaul flow variation step signals; apply the plurality of sets of input signals to the flow sensor and the expected ideal flow sensor model established in Step 2 respectively, to obtain actual flow response signals Q(k) and expected standard response signals Qc(k) respectively; the actual flow response signals Q(k) will pass through the established dynamic lag compensation unit Gc(z -1 ) to obtain compensated signals Qb(k); the fitness function F(k) of the genetic algorithm takes the absolute value of the difference between the compensated signals Qb(k) and the standard response signals Qc(k), and then the genetic algorithm continuously optimizes the parameters in the dynamic lag compensation unit Gc(z -1 ) in the direction of the minimum fitness function, and finally obtains the dynamic lag compensation unit Gc(z -1 ) so that the compensated signals Qb(k) continuously approach the expected standard response signals Qc(k); F(k) = |Qc(k) - Qb(k)| (4) where a c1 , a c2 , b c1 , and b c2 are model parameters of the dynamic lag compensator Gc(z -1 ); F(k) is a fitness function; Step 4, the model parameter initial value of the dynamic lag compensation unit Gc(z -1 ) established in step 3 will be determined through the initialization population step of the genetic algorithm, the upper and lower limits [a cmin ,a cmax ]、[b cmin ,b cmax ] of the Gc(z -1 ) parameter will be determined according to the parameters of the overall model, and the population number g, the crossover probability p c , the mutation probability p b of the genetic algorithm are set, the fitness function is shown in formula (4), and the three termination conditions are that the maximum genetic generation is Dmax, the best chromosome does not change continuously for n times, and the variance of the population fitness value is < the preset adjacent variance qj. Step 5, establishing a verification data set containing but not limited to a sine wave signal flow, and giving an actual flow sensor output Q(k) to a dynamic lag compensation unit to obtain a compensated signal Qb(k), and comparing the compensated signal Qb(k) with a given sensor dynamic performance response requirement, if the compensated signal Qb(k) satisfies the requirement, no further optimization is needed, and a final dynamic lag compensation unit is obtained; if the compensated signal Qb(k) does not satisfy the requirement, an identification data set is expanded, and the dynamic lag compensation unit is re-corrected until the dynamic lag compensation unit satisfies the given response performance requirement; Step 6, reasonably switching between an actual flow sensor output flow and a lag compensation output through a mutual correction algorithm to improve a response performance of the micro flow sensor.

2. The method of dynamic hysteresis compensation for a micro flow sensor of claim 1, wherein: An implementation method of step 6 is Specific steps of the mutual correction algorithm are as follows: Step 6.1, comparing whether the lag compensation flow Qb and the expected set flow Qset satisfy formula (5), Qb ≥ k1·Qset (5) If the formula (5) is satisfied, step 6.2 is performed; if the formula (5) is not satisfied, a mutual correction algorithm output correction flow Qy is equal to the lag compensation flow Qb; Qy = Qb, Qb < k1·Qset (6) In the formula, k1 is a response judgment coefficient, and a value range of k1 is 63%, 95% and 99%; Step 6.2, further comparing the lag compensation flow Qb and the actual flow response Q, if an absolute value of a difference between the lag compensation flow Qb and the actual flow response Q is less than a preset threshold value ε, step 6.4 is entered; if the absolute value of the difference between the lag compensation flow Qb and the actual flow response Q is greater than the preset threshold value ε, a mutual correction algorithm output correction flow Qy is equal to the lag compensation flow Qb: Qy = Qb, |Qb - Q| > ε (7); Step 6.3, the threshold value in step 6.2 is linearly fitted through maximum prediction errors of the lag compensation flow Qb and the actual flow response Q in a steady state under different step responses, and it is found through an empirical formula (8) that the errors are usually less than 3% of the actual flow response Q in the steady state, and therefore the threshold value ε is calculated through a public formula (9); e max = k2• U (8) ε = k3·Q (9) In the formula, e max is the maximum prediction error, U is the input signal value of the flow control valve, k2 is a fitting coefficient, k3 is a threshold correlation coefficient, and the value range is 5%, 3%, and 1%. Step 6.4, when conditions of steps 6.1 and 6.2 are satisfied, a mutual correction algorithm output correction flow Qy is equal to the actual flow response Q; Qy = Q, Qb ≥ k1·Qset and |Qb - Q| < ε (10) When switching is performed, if a condition in step 6.2 is continuously judged, an algorithm calculation capacity is affected, and a flow sensor output is caused to be dithered due to noise; if the condition is only judged once, a misjudgment is caused due to flow sensor noise; Therefore, a multiple judgment switching method in step 6.5 or a delay dithering judgment method is adopted. Step 6.5, multiple times of judgment switching: the absolute value of the difference between the hysteresis compensation flow Qb and the actual flow response Q calculated in 5 consecutive discrete periods T is within the threshold ε, then switching is performed, and after that, no longer judgment; delay anti-jitter judgment: after one judgment enters the threshold ε, delay for a period of time t to judge again whether there is one period in the next 3 discrete periods whose absolute value is within the threshold ε, if so, then switching is performed.

3. A dynamic hysteresis compensation system for a micro flow sensor for implementing a dynamic hysteresis compensation method for a micro flow sensor as claimed in claim 1 or 2, characterized in that: The gas source, the flow sensor, the microprocessor, and the flow control valve are connected in series, and the output interface of the flow sensor is connected to the input interface of the microprocessor, and the output interface of the microprocessor is connected to the input interface of the flow control valve, so as to adjust the opening size of the flow control valve.

4. The dynamic hysteresis compensation system of a micro flow sensor according to claim 3, characterized in that: the flow control valve is an execution unit of flow regulation; the flow sensor is used for collecting actual flow data output by the flow control valve; the microprocessor includes a dynamic hysteresis compensation unit, a mutual correction algorithm, and microprocessor calculation; the dynamic hysteresis compensation unit is used for compensating the hysteresis in the actual flow signal output by the flow sensor to obtain a hysteresis compensation flow; the mutual correction algorithm is used for reasonably switching between the actual flow sensor output flow and the hysteresis compensation output, so as to ensure responsiveness and accuracy; the microprocessor performs processing calculation according to the dynamic hysteresis compensation unit and the mutual correction algorithm, and sets the size of the flow control valve output flow.

Images