RBDE (Radial Basis Decision Element) algorithm-based robust anti-interference damping control method for wind turbine airfoil
Through the robust anti-interference and vibration-absorbing control method of wind airfoil based on RBDE algorithm, the problem of poor vibration control effect and insufficient robustness of wind airfoil under large-scale continuous gust interference is solved, and better anti-interference ability and robustness are achieved.
Patent Information
- Application Number
- CN202411466190.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-21
- Publication Date
- 2025-03-11
AI Technical Summary
The prior art is difficult to effectively control the vibration of the wind airfoil under large-scale continuous gust interference, resulting in poor control effect and insufficient robustness.
The robust anti-stop vibration damping control method of wind airfoil based on RBDE algorithm is adopted. By establishing a nonlinear vibration model, designing a vibration damping controller and driving compensator, and optimizing control parameters using the RBDE algorithm to improve immunity and robustness.
The fatigue load of airfoil vibration is significantly reduced, the driving saturation problem is overcome, and the robustness of wind airfoil anti-stop vibration damping control is enhanced under large-scale continuous gusts.
Smart Images
Figure CN119596681A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a wind turbine vibration reduction control technology, and in particular to a wind turbine airfoil robust anti-disturbance vibration reduction control method based on a RBDE algorithm. Background Art
[0002] With the development of large wind turbines, wind turbine blades are prone to vibration under the action of aeroelasticity, causing blade damage, thus threatening the safe and stable operation of wind turbines. Therefore, vibration reduction control of wind turbine airfoils is needed. However, the vibration of wind turbine airfoils is caused by the incoming wind speed. Sudden changes in the incoming wind speed, such as gusts, not only bring uncertain wind speed changes, but also bring great wind speed interference to the system, which seriously affects the vibration reduction control effect of wind turbine airfoils. In practice, there is a large range of continuous gust interference, and it is necessary to design an effective airfoil robust anti-disturbance vibration reduction control method to ensure the vibration reduction effect and robustness of wind turbine blades under wind conditions.
[0003] Linear active disturbance rejection control (LADRC) has the advantages of simple structure, estimation of unknown disturbances and no need for accurate models. LADRC based on extended state observer has strong anti-disturbance ability. At present, the existing LADRC literature has been widely successfully applied to energy and power systems. In recent years, it has also been applied to wind power generation systems. However, the existing wind power anti-disturbance control research is mainly aimed at generator systems and power grid systems, and there are few studies on the anti-disturbance control of wind turbine airfoil vibration systems. At the same time, the wind turbine airfoil vibration reduction control system is very sensitive to wind speed changes and disturbances. In order to improve the robustness of the control, it is necessary to adjust the optimal control parameters. However, although traditional intelligent optimization algorithms, such as differential evolution (DE) algorithms, have achieved good results in the calculation of the optimal control parameters of wind turbine airfoils, they cannot ensure the robustness of the closed-loop system under continuous wind disturbances. In addition, the drive saturation problem of the airfoil control surface is serious. Therefore, it is difficult to use existing methods to ensure excellent wind turbine airfoil anti-disturbance vibration reduction control effect and robustness under large-scale continuous gust disturbances. Summary of the invention
[0004] Purpose of the invention: In view of the above problems, the purpose of the present invention is to provide a robust anti-disturbance and vibration reduction control method for wind turbine airfoils based on the RBDE algorithm, which effectively improves the anti-disturbance ability, vibration reduction control effect, drive compensation ability and robustness of the wind turbine airfoil under a large range of continuous gusts.
[0005] Technical solution: The wind turbine airfoil robust anti-disturbance vibration reduction control method based on RBDE algorithm of the present invention comprises:
[0006] Using the aeroelastic model of wind turbine airfoils and wind tunnel test data, a nonlinear vibration model of wind turbine airfoils based on micro control surfaces is established.
[0007] Using the two-degree-of-freedom internal model control method, the vibration reduction controller K(s) of the wind turbine airfoil is constructed;
[0008] Based on the vibration reduction controller K(s) and LADRC control method, the vibration reduction controller G of the wind turbine airfoil is designed. c (s);
[0009] Design drive compensator u τ , combined with the vibration reduction controller G c (s), and obtain the controller g of the wind turbine airfoil;
[0010] The RBDE algorithm is used to optimize the control parameters in the controller g.
[0011] Furthermore, the expression of the nonlinear vibration model of the wind turbine airfoil is:
[0012]
[0013] Where M x is the mass matrix, C x is the damping matrix, K x is the stiffness matrix, F a is the aerodynamic matrix, x = [h α] T is the airfoil vibration vector, h is the airfoil flapping displacement, α is the airfoil torsion angle, is the first-order derivative of x, is the second-order derivative of x;
[0014] Among them, the expression of the aerodynamic matrix is:
[0015]
[0016] Where, L a 、M a are the aerodynamic lift and moment of the airfoil, ΔL g , ΔM g are the aerodynamic lift and moment of the micro control surface respectively.
[0017] Furthermore, the expression of the vibration reduction controller K(s) is:
[0018]
[0019] In the formula, s is a complex variable, Q c (s) is the internal model controller, Q d (s) is the disturbance rejection controller, P(s) is the system process model, and the expressions are:
[0020]
[0021] Where T is the time constant, ζ is the damping coefficient, k 0 is the system gain, λ is the feedback filter parameter, d is the anti-disturbance filtering parameter, b 1 is the zero point coefficient.
[0022] Furthermore, the vibration reduction controller G c The expression of (s) is:
[0023]
[0024] In the formula, k 1 is the control gain, β 1 is the feedback gain, β 2 is the observer gain, b is the adjustment factor;
[0025] The vibration control G c (s) is equivalent to the vibration reduction controller K(s), and the parameter k is obtained 1 , β 1 , β 2 Respectively expressed as:
[0026]
[0027] β 2 (λ d )=(1 / λ d ) 2 .
[0028] Furthermore, the drive compensator u τ The expression is:
[0029]
[0030] In the formula, t is time, k is p is the compensation gain, e(t) is the output feedback signal, is the speed feedback signal, is the acceleration feedback signal;
[0031] Then the expression of the controller g of the wind turbine airfoil is:
[0032]
[0033] Where α is the airfoil twist angle, represents the estimated value of the torsion angle, represents the estimated value of the torsion velocity, τ is the height of the micro-control surface, f is the unknown disturbance, is the estimated value of f, is the estimated value of the derivative of f, u is the control signal, k p , λ, λ d , b are the parameters to be optimized.
[0034] Furthermore, the optimization and solution of the control parameters in the controller g using the RBDE algorithm includes the following steps:
[0035] Step 51, set the parameter k to be optimized p , λ, λ d , b is the individual vector x of the RBDE algorithm i =[k pi ,λ i ,λ di ,b i ], i = 1, 2, ..., m, set the population size m and the maximum number of iterations G, and randomly generate the initial population;
[0036] Step 52, using the objective function J as the fitness function, calculating the fitness value of the individual vector, and selecting the optimal individual vector;
[0037] Among them, the objective function J and the constraint condition st are:
[0038]
[0039] Where ITAE is the integral of the relative error function, is the flapping speed of the airfoil, |u| is the absolute value of the control signal, is the absolute value of the control speed signal, g rmax is the driving speed limit, g max is the driving amplitude limit;
[0040] Step 53, update the closed-loop sensitivity ε of the optimal individual vector, expressed as:
[0041]
[0042] in,
[0043]
[0044] Where ω is the frequency, M s is the sensitivity to medium and low frequency interference, M p is the sensitivity to medium and high frequency disturbances, sup is the upper bound, max is the maximum value, and C(s) is the controller g;
[0045] Step 54, design a sensitivity adaptive mechanism according to the closed-loop sensitivity ε, and update the variation factor F and the crossover factor CR, the expressions are:
[0046]
[0047] In the formula, f m is the upper bound of the mutation factor, rand is a random number between 0 and 1, ε k is the closed-loop sensitivity of the kth generation optimal individual vector, cr m is the upper bound of the crossover factor;
[0048] Step 55, using the mutation factor F and the crossover factor CR to control the mutation operation, crossover operation and selection operation of the RBDE algorithm, repeating steps 51 to 54 until the iteration termination condition is met and the iteration is terminated to obtain the optimal individual vector.
[0049] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:
[0050] 1. The present invention proposes a RBDE algorithm that combines closed-loop sensitivity and differential evolution algorithm, which improves the global optimization performance of optimal control parameters compared with the traditional DE algorithm;
[0051] 2. Compared with the existing LADRC anti-disturbance control method, the method of the present invention significantly reduces the fatigue load of airfoil vibration for wind turbine airfoil anti-disturbance control under wind disturbance, and effectively overcomes the problem of drive saturation; 3. Under large-scale continuous gusts, the present invention enhances the robustness of wind turbine airfoil anti-disturbance vibration control. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 It is a flow chart of the robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm;
[0053] Figure 2 It is the flowchart of RBDE algorithm;
[0054] Figure 3 This is a comparison chart of the objective function calculation results based on different evolutionary algorithms;
[0055] Figure 4 It is a picture of sustained gusts over a large area;
[0056] Figure 5 The anti-disturbance vibration reduction control response of wind turbine airfoil under large-scale continuous gusts. DETAILED DESCRIPTION
[0057] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments.
[0058] The robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm described in this embodiment is as follows: Figure 1 As shown, the method comprises the following steps:
[0059] Step 1: Using the aeroelastic model of wind turbine airfoil and wind tunnel test data, a nonlinear vibration model of wind turbine airfoil based on micro control surfaces is established.
[0060] Among them, the wind tunnel test data include the aerodynamic lift coefficient and moment coefficient of the wind turbine airfoil, as well as the aerodynamic lift coefficient and moment coefficient of the micro control surface.
[0061] Step 2: Use the two-degree-of-freedom internal model control method to construct the vibration reduction controller K(s) of the wind turbine airfoil.
[0062] Step 3: Based on the vibration reduction controller K(s) and the LADRC control method, the vibration reduction controller G of the wind turbine airfoil is designed. c (s).
[0063] Among them, the vibration reduction controller G c (s) Has anti-interference capability.
[0064] Step 4: Design the drive compensator u τ , combined with the vibration reduction controller G c (s), and obtain the controller g of the wind turbine airfoil.
[0065] Among them, the controller g serves as the final vibration reduction controller of the wind turbine airfoil.
[0066] Step 5: Use the RBDE algorithm to optimize and solve the control parameters in the controller g.
[0067] Among them, the RBDE (sensitivity differential evolution) algorithm is a new differential evolution algorithm proposed by the present invention, which is based on the closed-loop system sensitivity design. The RBDE algorithm is used to solve and optimize the unknown parameters in the controller g, thereby realizing the robust anti-disturbance vibration reduction control of the wind turbine airfoil.
[0068] The present invention aims at the vibration reduction problem of wind turbine airfoils under large-scale continuous gust interference, and provides a robust anti-disturbance vibration reduction control method for wind turbine airfoils based on the RBDE algorithm; aims at the drive saturation problem of airfoil control surfaces, and designs a new drive compensator; aims at the optimal parameter setting problem of the robust anti-disturbance vibration reduction controller, and designs a sensitivity differential evolution algorithm, namely the RBDE algorithm. Compared with the traditional DE algorithm, the RBDE algorithm of the present invention significantly improves the global search capability of the optimal control parameters, and improves the effectiveness and robustness of the set parameters.
[0069] In one example, the expression of the nonlinear vibration model of the wind turbine airfoil in step 1 is:
[0070]
[0071] Where M x is the mass matrix, Cx is the damping matrix, K x is the stiffness matrix, F a is the aerodynamic matrix, x = [h α] T is the airfoil vibration vector, h is the airfoil flapping displacement, α is the airfoil torsion angle, is the first-order derivative of x, is the second-order derivative of x;
[0072] Among them, the expression of the aerodynamic matrix is:
[0073]
[0074] Where, L a 、M a are the aerodynamic lift and moment of the airfoil, ΔL g , ΔM g are the aerodynamic lift and moment of the micro control surface respectively. a 、M a and ΔL g , ΔM g The expressions are:
[0075]
[0076] In the formula, ρ is air density, U is wind speed, b s is the half chord length of the airfoil, a is the distance coefficient, c l,α 、c m,α are the aerodynamic lift coefficient and moment coefficient of the airfoil, c l,g 、c m,g is the aerodynamic lift coefficient and moment coefficient of the micro control surface; c l,α ,c m,α and c l,g ,c m,g Obtained through airfoil wind tunnel test data.
[0077] In one example, the damping controller K(s) in step 2 is expressed as:
[0078]
[0079] In the formula, s is a complex variable, Q c (s) is the internal model controller, Q d (s) is the disturbance rejection controller, P(s) is the system process model, and the expressions are:
[0080]
[0081] Where T is the time constant, ζ is the damping coefficient, k 0 is the system gain, λ is the feedback filter parameter, dis the anti-disturbance filtering parameter, b 1 is the zero point coefficient.
[0082] Q c (s), Q d Substituting the formula of P(s) and P(s) into the expression of K(s), we get:
[0083]
[0084] The equivalent form of K(s) in the above formula is PID controller and integrator:
[0085]
[0086] Furthermore, the vibration reduction controller G c The expression of (s) is:
[0087]
[0088] In the formula, k 1 is the control gain, β 1 is the feedback gain, β 2 is the observer gain, b is the adjustment factor;
[0089] In the above formula, G c (s) is equivalent to the PID controller and integrator in the form of:
[0090]
[0091] The vibration control G c (s) is equivalent to the vibration reduction controller K(s), then Find the parameter k 1 , β 1 , β 2 They are:
[0092]
[0093] β 2 (λ d )=(1 / λ d ) 2 .
[0094] Furthermore, the drive compensator u τ The expression is:
[0095]
[0096] In the formula, t is time, k is p is the compensation gain, e(t) is the output feedback signal, is the speed feedback signal, is the acceleration feedback signal;
[0097] Then the expression of the controller g of the wind turbine airfoil is:
[0098]
[0099] Where α is the airfoil twist angle, represents the estimated value of the torsion angle, represents the estimated value of the torsion velocity, τ is the height of the micro-control surface, f is the unknown disturbance, is the estimated value of f, is the estimated value of the derivative of f, u is the control signal, k p , λ, λ d , b are the parameters to be optimized.
[0100] Furthermore, the optimization and solution of the control parameters in the controller g using the RBDE algorithm includes the following steps:
[0101] Step 51, set the parameter k to be optimized p , λ, λ d , b is the individual vector x of the RBDE algorithm i =[k pi ,λ i ,λ di ,b i ], i = 1, 2, ..., m, set the population size m and the maximum number of iterations G, and randomly generate the initial population;
[0102] Step 52, using the objective function J as the fitness function, calculating the fitness value of the individual vector, and selecting the optimal individual vector;
[0103] Among them, the objective function J and the constraint condition st are:
[0104]
[0105] Where ITAE is the integral of the relative error function, is the flapping speed of the airfoil, |u| is the absolute value of the control signal, is the absolute value of the control speed signal, g rmax is the driving speed limit, g max is the driving amplitude limit;
[0106] Step 53, update the closed-loop sensitivity ε of the optimal individual vector, expressed as:
[0107]
[0108] in,
[0109]
[0110] Where ω is the frequency, M s is the sensitivity to medium and low frequency interference, M p is the sensitivity to medium and high frequency disturbances, sup is the upper bound, max is the maximum value, and C(s) is the controller g;
[0111] Step 54, design a sensitivity adaptive mechanism according to the closed-loop sensitivity ε, and update the variation factor F and the crossover factor CR, the expressions are:
[0112]
[0113] In the formula, f m is the upper bound of the mutation factor, rand is a random number between 0 and 1, ε k is the closed-loop sensitivity of the kth generation optimal individual vector, cr m is the upper bound of the crossover factor;
[0114] Step 55, using the mutation factor F and the crossover factor CR to control the mutation operation, crossover operation and selection operation of the RBDE algorithm, repeating steps 51 to 54 until the iteration termination condition is met and the iteration is terminated to obtain the optimal individual vector.
[0115] Furthermore, the implementation process of step 55 includes:
[0116] The mutation factor F is used to control the mutation operation of the RBDE algorithm and obtain the mutation vector h i :
[0117] h i (k+1)=x p1 (k)+F(x p2 (k)-x p3 (k))
[0118] In the formula, x p1 ,x p2 ,x p3 are three random individual vectors in the population, and satisfy p1≠p2≠p3≠i;
[0119] The crossover operation of the RBDE algorithm is controlled by the crossover factor CR to obtain the experimental vector v i :
[0120]
[0121] In the formula, rand is a random number between 0 and 1;
[0122] Comparison experiment vector v i and individual vector x iThe fitness value of , select the vector with better fitness value as the next generation individual vector:
[0123]
[0124] When the number of iterations does not reach the maximum number of iterations G, return to step 51 and enter the next evolution cycle; when the number of iterations reaches the maximum number of iterations G, terminate the algorithm and output the optimal individual vector.
[0125] In order to verify the effectiveness of the method of the present invention, the following examples are used to further illustrate it.
[0126] For the DU97W300 wind turbine airfoil, the micro control surface is located at the tail end of the airfoil, and the driving limit of the micro control surface is g max =2%c,g rmax = 10% c, where c = 2b s is the chord length of the airfoil. The system-related parameters are ρ=1.225, b s =0.135, a=-0.6847, U=15. The parameters of the RBDE algorithm described in the present invention are set as follows: m=10, G=20, f m =1,cr m =1.2. The parameters of the traditional DE algorithm are set as follows: m=10, G=20, F=1, CR=0.2. The design process of the wind turbine airfoil robust anti-disturbance vibration reduction control based on the RBDE algorithm of the present invention is as follows: Figure 2 The objective function calculation results of the RBDE algorithm and the traditional DE algorithm are shown in Figure 3 As shown in FIG. 1 , FIG. (a) corresponds to the RBDE algorithm of the present invention, and FIG. (b) corresponds to the traditional DE algorithm. Figure 3 It can be seen that by comparing with the traditional DE algorithm, the RBDE algorithm of the present invention shows superiority in both the global search and local search of the optimal control parameters.
[0127] Widespread sustained gusty winds Figure 4 As shown, the wind speed range is [12,20] m / s, and there are gust disturbances at 6 seconds and 14 seconds. The robust anti-disturbance vibration reduction control results of wind turbine airfoil under a wide range of continuous gusts using the RBDE algorithm of the present invention, the classic LADRC method, and the LADRC method based on the traditional compensator are shown, where Figure (a) is a comparison of the flapping displacement response, Figure (b) is a comparison of the torsion angle response, and Figure (c) is a comparison of the control surface drive response. The control performance results of different controllers are shown in Table 1.
[0128] Table 1 Anti-disturbance and vibration reduction control performance indicators of wind turbine airfoil under wind disturbance
[0129]
[0130]
[0131] Figure 5 As can be seen from Table 1, by comparing the classic LADRC anti-disturbance control method and the LADRC method based on the traditional compensator, the method of the present invention reduces the torsional fatigue load by 98.9% and 99.1%, the flapping fatigue load by 97.5% and 96.1% under a large range of continuous gust disturbances, and reduces the drive strain by 97.4% and 96.7%, while overcoming the drive saturation problem that is difficult to solve with the existing LADRC method. In addition, for gust disturbances in different periods of time, the method of the present invention maintains an excellent wind turbine airfoil anti-disturbance and vibration reduction effect. Therefore, the method of the present invention, in comparison with the existing anti-disturbance control method, has the advantages of good anti-wind disturbance effect, good vibration reduction performance, strong drive compensation capability and robustness for the case of large-scale continuous gust disturbances.
Claims
1. A robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm, characterized in that: include: Using the aeroelastic model of wind turbine airfoils and wind tunnel test data, a nonlinear vibration model of wind turbine airfoils based on micro control surfaces is established. Using the two-degree-of-freedom internal model control method, the vibration reduction controller K(s) of the wind turbine airfoil is constructed; Based on the vibration reduction controller K(s) and LADRC control method, the vibration reduction controller Gc(s) of the wind turbine airfoil is designed; Design drive compensator u τ , combined with the vibration reduction controller Gc(s), the controller g of the wind turbine airfoil is obtained; The RBDE algorithm is used to optimize the control parameters in the controller g.
2. The robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm according to claim 1 is characterized in that: The expression of the nonlinear vibration model of the wind turbine airfoil is: Where M x is the mass matrix, C x is the damping matrix, K x is the stiffness matrix, F a is the aerodynamic matrix, x = [h α] T is the airfoil vibration vector, h is the airfoil flapping displacement, α is the airfoil torsion angle, is the first-order derivative of x, is the second-order derivative of x; Among them, the expression of the aerodynamic matrix is: Where, L a 、M a are the aerodynamic lift and moment of the airfoil, ΔL g , ΔM g are the aerodynamic lift and moment of the micro control surface respectively.
3. The robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm according to claim 2 is characterized in that: The expression of the vibration reduction controller K(s) is: In the formula, s is a complex variable, Q c (s) is the internal model controller, Q d (s) is the disturbance rejection controller, P(s) is the system process model, and the expressions are: Where T is the time constant, ζ is the damping coefficient, k0 is the system gain, λ is the feedback filter parameter, and d is the anti-disturbance filtering parameter, b1 is the zero point coefficient.
4. The robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm according to claim 3 is characterized in that: The expression of the vibration reduction controller Gc(s) is: In the formula, k1 is the control gain, β1 is the feedback gain, β2 is the observer gain, and b is the adjustment factor; The vibration control G c (s) is equivalent to the vibration reduction controller K(s), and the parameters k1, β1, and β2 are expressed as: β2(λ d )=(1 / l d ) 2 。 5. The robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm according to claim 4 is characterized in that: Drive compensator u τ The expression is: In the formula, t is time, k is p is the compensation gain, e(t) is the output feedback signal, is the speed feedback signal, is the acceleration feedback signal; Then the expression of the controller g of the wind turbine airfoil is: Where α is the airfoil twist angle, represents the estimated value of the torsion angle, represents the estimated value of the torsion velocity, τ is the height of the micro-control surface, f is the unknown disturbance, is the estimated value of f, is the estimated value of the derivative of f, u is the control signal, k p , λ, λ d , b are the parameters to be optimized.
6. The robust anti-disturbance vibration reduction control method for wind turbine airfoil based on RBDE algorithm according to claim 5 is characterized in that: The optimization and solution of the control parameters in the controller g using the RBDE algorithm includes the following steps: Step 51: Set the parameter k to be optimized p , λ, λ d , b is the individual vector x of the RBDE algorithm i =[k pi ,λ i ,λ di , b i ], i = 1, 2, ..., m, set the population size m and the maximum number of iterations G, and randomly generate the initial population; Step 52, using the objective function J as the fitness function, calculating the fitness value of the individual vector, and selecting the optimal individual vector; Among them, the objective function J and the constraint condition st are: Where ITAE is the integral of the relative error function, is the flapping speed of the airfoil, |u| is the absolute value of the control signal, is the absolute value of the control speed signal, g rmax is the driving speed limit, g max is the driving amplitude limit; Step 53, update the closed-loop sensitivity ε of the optimal individual vector, expressed as: in, Where ω is the frequency, M s is the sensitivity to medium and low frequency interference, M p is the sensitivity to medium and high frequency disturbances, sup is the upper bound, max is the maximum value, and C(s) is the controller g; Step 54, design a sensitivity adaptive mechanism according to the closed-loop sensitivity ε, and update the variation factor F and the crossover factor CR, the expressions are: In the formula, f m is the upper bound of the mutation factor, rand is a random number between 0 and 1, ε k is the closed-loop sensitivity of the kth generation optimal individual vector, cr m is the upper bound of the crossover factor; Step 55, using the mutation factor F and the crossover factor CR to control the mutation operation, crossover operation and selection operation of the RBDE algorithm, repeating steps 51 to 54 until the iteration termination condition is met and the iteration is terminated to obtain the optimal individual vector.