A time delay identification method for a large-scale transient high-speed wind tunnel

Abstract

The application discloses a time delay identification method for a large-scale transient high-speed wind tunnel, and relates to the field of wind tunnel test, which comprises the following steps: S1, a Wiener nonlinear model with dynamic characteristics is used to model a fence-stationary pressure system of the transient high-speed wind tunnel; S2, based on the collected data, a least square method is used to identify parameters of the Wiener nonlinear model to obtain an estimated value corresponding to the stationary pressure; S3, the method of S2 is used to estimate a plurality of groups of data, and based on the estimation result of S2, an estimated value corresponding to a model with the minimum estimation error is selected as a time delay of the system. After the time delay estimation value of the system is accurately obtained, a time delay compensation link can be added in the controller design, that is, the fence is operated in advance before the angle of attack changes, so that the influence of the time delay on the system control is eliminated.

Description

Technical Field

[0001] This invention relates to the field of wind tunnel testing. More specifically, this invention relates to a time delay identification method for large-scale transient high-speed wind tunnels. Background Technology

[0002] Wind tunnel systems are important ground-based simulation platforms for studying airflow patterns and the aerodynamic characteristics of aerospace vehicles and other objects. Essentially, they are tubular test devices capable of generating controllable, uniform airflow. Wind tunnel testing is a core means of acquiring aerodynamic data of aircraft, evaluating the rationality of layout design, and predicting flight performance. The accuracy of wind tunnel testing directly determines the optimization direction of aircraft layout schemes and the reliability of actual flight performance, thus having a crucial impact on the quality and technological level of aircraft development.

[0003] The structure of the 2.4M transient transonic wind tunnel (i.e., the large transient high-speed wind tunnel referred to in this patent) is as follows: Figure 4 As shown. A stable airflow environment is a prerequisite for ensuring the accuracy of wind tunnel tests; therefore, the design of the controller used to maintain flow stability in a wind tunnel system is crucial. However, the inherent large time delay characteristic of wind tunnel systems presents challenges to controller design. Figure 4 As can be seen, due to the difference in installation position between the grid finger system A (actuator) and the flow field measurement point B, the effect of the grid finger on the static pressure of the airflow field after its action requires a certain time delay before it is reflected in the detection data. Therefore, accurately obtaining this time delay parameter is of great significance for optimizing controller design and improving the accuracy of wind tunnel flow field control. Summary of the Invention

[0004] One object of the present invention is to solve at least the above-mentioned problems and / or defects, and to provide at least the advantages described below.

[0005] To achieve these objectives and other advantages of the present invention, a time-delay identification method for a large transient high-speed wind tunnel is provided, comprising: S1. The grid finger-static pressure system of the transient high-speed wind tunnel is modeled using the Wiener nonlinear model with dynamic characteristics. S2. Based on the data collected on site, the least squares method is used to identify the parameters of the Wiener nonlinear model in order to obtain the estimated value of the corresponding static pressure. S3. Estimate the data using the method of S2 for multiple sets of data, and based on the estimation results of S2, select the estimated value corresponding to the model with the smallest estimation error as the time delay of the system. In S1, the Wiener nonlinear model is expressed by the following equation: In the above formula, rLet be the order of the polynomial. c j are polynomial parameters, and c 1=1, n The order of the linear part of the model. a i and b i For model parameters, d The time delay order is... u t-d-i For gate finger control quantity, y t This refers to the static pressure output by the system.

[0006] Preferably, in S1, the construction process of the Wiener nonlinear model includes: S10. Construct the Wiener nonlinear initial model corresponding to the gate finger-static pressure system based on time delay, linear dynamic block, and nonlinear static block; S11. The linear dynamic block in the Wiener nonlinear initial model is parameterized using the following formula: In the above formula, v t-i For unmeasurable intermediate variables, i Sampling time; S12. Based on wind tunnel characteristics, the output nonlinearity in the Wiener nonlinear initial model is modeled using the following formula: In the above formula, F -1 It is the symbol for a nonlinear inverse function. Output basis functions for the system. j For exponential coefficients; S13. Based on the modifications of S11 and S12, a Wiener nonlinear model with dynamic characteristics is obtained.

[0007] Preferably, in S2, the static pressure estimate is obtained as follows: S20. Based on the data collected on-site, the model parameters are calculated using the following formula. a i , b i , c j Identify: In the above formula, the superscript T is the transpose symbol. θ Let be the parameter vector, and , F tFor information vectors, defined as ; S21. Define the output set Gate control quantity set ; ,in, L Given the data length, the parameter estimates in the Wiener nonlinear model are represented by the following formula: In the above formula, For parameter estimation vectors; S22, Estimated value of static pressure Characterized by the following formula: .

[0008] Preferably, in S3, the time delay is obtained in the following way: S30. Set a corresponding time delay estimate for any set of data. d =0, iteration count D =20; S31. For any set of data, the estimation error of the Wiener nonlinear model pair is obtained using the following formula. E : In the above formula, This is an estimated value for static pressure. The output information vector is defined as follows: , L For data length; S32, when d < D season d = d +1, return to S2; when d = D At this point, the estimation iteration for the corresponding data ends; S33. For any set of data, select the model with the smallest estimation error from the average estimates of the corresponding MSE. d The value is used as the time delay of the system.

[0009] The present invention has at least the following beneficial effects: After accurately obtaining the estimated time delay of the system, this invention can add a time delay compensation step to the controller design, that is, operate the gate finger in advance before the angle of attack changes, thereby eliminating the influence of time delay on system control.

[0010] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the grid finger-static pressure system. Figure 2 This is a schematic diagram of the Wiener system model of the present invention; Figure 3 This is a schematic diagram of the identification error results using the Wiener system model in this invention; Figure 4 This is a schematic diagram of the equipment layout corresponding to the static pressure model in a 2.4M transient transonic wind tunnel; Figure 5 The diagram shows the control effect using the original controller; Figure 6 The diagram shows the control effect of the controller after applying the time delay compensation of the present invention. Detailed Implementation

[0012] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.

[0013] This invention provides a time delay identification method for a large transient high-speed wind tunnel. It is a time delay estimation method mainly based on model identification. This method equates the grid finger-static pressure system to a Wiener nonlinear model (hereinafter referred to as the W model). Combining data obtained from conventional wind tunnel tests, the model estimation error under different time delay assumptions is solved through the model identification process. Finally, the true time delay of the system is determined based on the error characteristics. Reliable time delay estimation results of the static pressure measurement link of the 2.4M transient transonic wind tunnel system are obtained, providing key parameter support for the subsequent controller optimization design.

[0014] Specifically, the time delay estimation method based on model identification of this invention includes the following: System time delay estimation based on model identification first requires establishing a mathematical model of the system, and then estimating the system's time delay based on the modeling results. Therefore, the establishment of a static pressure model is studied first. The static pressure in the wind tunnel airflow field is mainly controlled by a grid finger system, the composition of which is as follows: Figure 1 As shown, it mainly includes a grating system and an airflow field. In application, the grating control quantity input to the system is converted into the corresponding grating displacement through the grating system, and then outputs the corresponding static pressure value after being acted upon by the airflow field.

[0015] Typically, the grating finger system is driven by a hydraulic system, and a typical hydraulic system can be modeled using a linear dynamic model. For the airflow field, a nonlinear model with dynamic characteristics can be used. Considering the hysteresis characteristics present in the system, the model of the grating finger-static pressure system can be derived as follows: Figure 2 The W model shown is used to represent this, where, Figure 2 Iny t This is the system output, i.e., static pressure (unit: Pa). u t The gate finger control quantity input to the system (unit: mm); v t It is an intermediate variable that cannot be measured; z -1 For sampling shift operator; G ( u t ) is a linear dynamic block characteristic function. F ( v t The output is a nonlinear function. It should be noted that in the W model, this is achieved through a sampling shift operator. z -1 To characterize the gate finger control quantity of the system input u t The time delay is used to obtain the corresponding linear dynamic block characteristic function. G ( u t Further, through unmeasurable intermediate variables; v t Will G ( u t Convert to a nonlinear static function F ( v t ), to obtain the system output y t ; The W model is then parametrically represented, and the characteristics of the linear dynamic block in the model are described as follows: (1) in, n The order of the linear part of the model; a i and b i These are model parameters; d The time delay order is... v t-i For unmeasurable intermediate variables, i The sampling time.

[0016] According to wind tunnel characteristics, when the grating finger position remains unchanged, the airflow field will generate a unique static pressure value, and vice versa. That is, the relationship between grating finger and static pressure is one-to-one. Therefore, the inverse characteristics of the nonlinear output can be modeled. This invention selects the following polynomial function to represent the pseudo-characteristics of the nonlinear static part of the model: (2) In the above formula,r Let be the order of the polynomial. F -1 It is the symbol for a nonlinear inverse function. Output basis functions for the system. j For exponential coefficients, c j Let be the polynomial parameter, where c 1 = 1.

[0017] Based on equations (1) and (2), the parameterized W model is obtained: (3).

[0018] This invention uses conventional experimental data, sets different time delays, and identifies parameters based on equation (3) to obtain an estimate of the static pressure. Based on the estimation results, the model with the smallest estimation error is selected. d The value is used as an estimate of the time delay. Specifically, the estimation algorithm steps are as follows: Step 1: Settings d =0; Step 2: Identify model parameters using data collected on-site. a i , b i and c j ; In step 2, the model parameters are identified using the following least squares method: (4) in, θ Let the parameter vector be defined as follows: , F t For information vectors, defined as .

[0019] definition , , ,in L This represents the data length.

[0020] Therefore, the parameter estimates in equation (4) are: (5) In the above formula, This is the parameter estimation vector.

[0021] Step 3: Estimate the system output static pressure value based on the parameter identification results; In step 3, the static pressure is estimated by the following formula: (6).

[0022] Step 4: Calculate the model estimation error E ; Regarding the error in step 4 E The mean square error is used for calculation, that is: (7) in, This is an estimated value for static pressure. The output information vector is defined as follows: , L For data length; Step 5: Settings D =20, when d < D season d = d +1, return to step 1; when d = D The calculation ends when the time is reached.

[0023] Taking the atmospheric pressure test under the 0.8 Ma condition as an example, the time delay is estimated. In the estimation process, 30 sets of data are selected for calculation of algorithm steps 1 to 5 respectively.

[0024] For all d The value is the estimated error obtained from 30 sets of data. E In practical applications, estimation error E Adopting such Figure 3 The mean square error (MSE) shown is used to characterize this, where, Figure 3 The middle red line is a certain d The value corresponds to any of the 30 MSEs. d The average of the 30 MSEs corresponding to the value (shown by the blue asterisks in the graph), i.e., the red line, is calculated using the following formula: (8) in, MSE i To target a specific d Value, using the first i The mean square error obtained from the identification of the set of data.

[0025] pass Figure 3 The identification results show that when d When =7, M MSE The minimum value is obtained, therefore, under the test condition of 0.8 Ma atmospheric pressure, the time delay of the system is considered to be... d =7.

[0026] The time delay identification method described above can be used to obtain reliable time delay parameters of large transient wind tunnel systems, providing key data support for subsequent engineering applications such as system output prediction, time delay compensation, and control strategy optimization.

[0027] Verification example: Taking the test condition of 0.825mA and normal pressure as an example, the control effect obtained by the conventional controller is as follows: Figure 5 As shown, the control effect using time delay compensation is as follows: Figure 6 As shown in the figure, the control effect of the two is significantly improved and the accuracy is greatly improved after adding time delay compensation, which is sufficient to demonstrate the effectiveness of the present invention.

[0028] The above solution is merely an illustration of a preferred example and is not limited thereto. When implementing this invention, appropriate substitutions and / or modifications can be made according to the user's needs.

[0029] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. It can be applied to various fields suitable for the present invention. Other modifications can be readily made by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and examples shown and described herein.

Claims

1. A method for time delay identification in a large transient high-speed wind tunnel, characterized in that, include: S1. The grid finger-static pressure system of the transient high-speed wind tunnel is modeled using the Wiener nonlinear model with dynamic characteristics. S2. Based on the data collected on site, the least squares method is used to identify the parameters of the Wiener nonlinear model in order to obtain the estimated value of the corresponding static pressure. S3. Estimate the data using the method of S2 for multiple sets of data, and based on the estimation results of S2, select the estimated value corresponding to the model with the smallest estimation error as the time delay of the system. In S1, the Wiener nonlinear model is expressed by the following equation: In the above formula, r Let be the order of the polynomial. c j are polynomial parameters, and c 1=1, n The order of the linear part of the model. a i and b i For model parameters, d The time delay order is... u t-d-i For gate finger control quantity, y t The static pressure output by the system; In S3, the time delay is obtained as follows: S30. Set a corresponding time delay estimate for any set of data. d =0, iteration count D =20; S31. For any set of data, the estimation error of the Wiener nonlinear model pair is obtained using the following formula. E : In the above formula, This is an estimated value for static pressure. The output information vector is defined as follows: , L For data length; S32, when d < D season d = d +1, return to S2; when d = D At this point, the estimation iteration for the corresponding data ends; S33. For any set of data, select the model with the smallest estimation error from the average estimates of the corresponding MSE. d The value is used as the time delay of the system.

2. The time delay identification method for a large transient high-speed wind tunnel as described in claim 1, characterized in that, In S1, the construction process of the Wiener nonlinear model includes: S10. Construct the Wiener nonlinear initial model corresponding to the gate finger-static pressure system based on time delay, linear dynamic block, and nonlinear static block; S11. The linear dynamic block in the Wiener nonlinear initial model is parameterized using the following formula: In the above formula, v t-i For unmeasurable intermediate variables, i Sampling time; S12. Based on wind tunnel characteristics, the output nonlinearity in the Wiener nonlinear initial model is modeled using the following formula: In the above formula, F -1 It is the symbol for a nonlinear inverse function. Output basis functions for the system. j For exponential coefficients; S13. Based on the modifications of S11 and S12, a Wiener nonlinear model with dynamic characteristics is obtained.

3. The time delay identification method for a large transient high-speed wind tunnel as described in claim 1, characterized in that, In S2, the static pressure estimate is obtained as follows: S20. Based on the data collected on-site, the model parameters are calculated using the following formula. a i , b i , c j Identify: In the above formula, the superscript T is the transpose symbol. θ Let be the parameter vector, and , F t For information vectors, defined as S21. Define the output set Gate control quantity set ; ,in, L Given the data length, the parameter estimates in the Wiener nonlinear model are represented by the following formula: In the above formula, For parameter estimation vectors; S22, Estimated value of static pressure Characterized by the following formula: 。

Images