Disclosure of Invention
The invention aims to: aiming at the problem that the existing feeding system servo optimization method cannot realize simultaneous optimization of multiple aspects of the performance of a servo system and real-time optimization on demand, the invention provides a feeding system servo optimization method based on an extensible dynamic performance evaluation function.
The technical scheme is as follows: the invention discloses a feeding system servo optimization method based on a scalable dynamic performance evaluation function, which comprises the following steps of:
(1) Establishing a servo simulation model of a feeding system based on MATLAB;
(2) Determining a servo parameter optimization variable of a feed drive;
(3) Constructing a servo-optimized extensible dynamic performance evaluation function of the feeding system, selecting respective corresponding response characteristic index evaluation functions according to certain dynamic response performances to be comprehensively considered by the feeding system, combining actual optimization requirements of the feeding system, and substituting the functions into the extensible dynamic performance evaluation function to obtain a servo-optimized target evaluation function;
(4) Based on an intelligent optimization algorithm, the performance of the feeding system is optimized by servo parameter optimization in the direction of the constructed target evaluation function.
By constructing the scalable dynamic performance evaluation function of the servo optimization, the real-time adjustment and the further integration of the servo optimization target evaluation function can be conveniently realized on the basis of the scalability of the dynamic response characteristic index evaluation function according to the actual optimization requirement, so that the performance of the servo system can be conveniently optimized as required.
In the step (2), the servo parameter optimization variables include position loop gain, velocity loop gain, current loop gain, and integration time.
In the step (3), the expression of the scalable dynamic performance evaluation function SOEF for the servo optimization of the feeding system is as follows:
in the formula, the performance evaluation index SPI of the feeding system is a response characteristic index evaluation function, n is the number of the performance evaluation indexes SPI of the system, and w is the number of the performance evaluation indexes SPI of the system i Scalability factor in a scalable dynamic performance evaluation function for a single performance evaluation index, w i Has a value of 0 or 1,f i For system performance evaluation index SPI i The influence coefficient of (c).
Specifically, when a performance evaluation index SPI does not affect the scalable dynamic performance evaluation function SOEF, w i =0, i.e. the performance evaluation index is not taken into account in the optimization of the response characteristics of the feed system; when a certain performance evaluation index SPI has an influence on the scalable dynamic performance evaluation function SOEF, w i =1, i.e. the performance evaluation index is taken into account in the optimization of the response characteristics of the feed system; when w is i When =1, by f i To further control the SPI i The degree of influence on the overall evaluation function. f. of i The value of (b) can be any positive number, and the value of (b) can be selected to be a positive integer in general for convenient calculation; f. of i The larger the value is, the performance evaluation index SPI is represented i The greater the influence on the overall evaluation function, so it is correctBody evaluation function or dynamic performance optimization requirement emphasizes certain performance evaluation index SPI i The corresponding system performance evaluation index influence coefficient f i Larger values may be taken.
In the step (3), the dynamic response performance of the feeding system may include overshoot and transient oscillation response performance, and the corresponding response characteristic index evaluation function is preferably:
in the above formula, phi s To displace a set value, phi a Is the actual value of the displacement; a represents overshoot balance coefficient, b represents transient oscillation balance coefficient, and the values of a and b are arbitrary positive numbers p 0 Is in the form of p 1 t is a linear function or p of the time variable t 2 t 2 With respect to the time variable t, where p 1 Coefficient representing a linear function of time, p 2 Coefficients representing a quadratic time function; p is a radical of 0 、p 1 、p 2 The value is positive, and usually, for convenience of calculation, the value is selected to be a positive integer.
The overshoot trade-off coefficient a and the transient oscillation trade-off coefficient b have the following priorities: preference for positive number p 0 Then, choose the shape as p 1 t as a function of time, and finally selecting the form p 2 t 2 Is a quadratic function of time. When the overshoot phenomenon is emphasized more on the optimization requirement of the overshoot and transient oscillation response performance, the priority of the value of the corresponding overshoot weighing coefficient a is higher than that of the transient oscillation weighing coefficient b, and the values of a and b meet p 0 >p 1 >p 2 (ii) a For example, if a takes on a positive number p 0 The value of b is selected to be p 1 A time function of t or the form p 2 t 2 Of a quadratic time function of, correspondingly, p 0 >p 1 Or p 0 >p 2 (ii) a If a takes the form of p 1 t as a function of time, and b is chosen to be p 2 t 2 Of a quadratic time function of, while p 1 >p 2 . On the contrary, when the transient oscillation phenomenon is emphasized more on the optimization requirement of the overshoot and transient oscillation response performance, the priority of the value of the corresponding transient oscillation weighing coefficient b is higher than that of the overshoot weighing coefficient a, and the values of a and b meet p 0 >p 1 >p 2 . After the value types (positive number, primary function or quadratic function) of a and b are determined according to optimization requirements and value priorities, corresponding p can be processed according to actual optimization conditions 0 、p 1 Or p 2 Is adjusted by the value of p 0 、p 1 Or p 2 The larger the value of (a) is, the larger the value of the corresponding coefficient a or b is, indicating that the overshoot corresponding to the coefficient a or the transient oscillation corresponding to the coefficient b has a greater influence on the dynamic performance optimization result at the selected priority, and vice versa.
When the overshoot and transient oscillation response performance is optimized, and the importance degrees of the optimization requirements of the overshoot phenomenon and the transient oscillation phenomenon are the same or similar, the values of the corresponding overshoot weighing coefficient a and the transient oscillation weighing coefficient b are the same priority. For example, a and b may be shaped as p at the same time 2 t 2 The quadratic time function of (2) can also be taken simultaneously as p 1 A time function of t or a positive number p 0 。
Has the advantages that: compared with the prior servo optimization method of the feeding system, the invention has the beneficial effects that: the method for optimizing the servo of the feeding system can simultaneously realize the control effect optimization of the system in multiple aspects based on the scalability of the objective function by constructing the scalable dynamic performance evaluation function of the servo optimization according to the actual optimization requirement, and can conveniently realize the real-time adjustment and further integration of the objective function of the servo optimization of the feeding system, thereby conveniently realizing the performance optimization of the servo system as required.
Detailed Description
The technical scheme of the invention is further explained by combining the attached drawings.
According to the feeding system servo optimization method based on the scalable dynamic performance evaluation function, the scalable target evaluation function is constructed through servo optimization, the scalability of the target function is evaluated based on the response characteristic, and the real-time adjustment and the further integration of the servo optimization target function can be conveniently realized according to the actual optimization requirement, so that the performance of the servo system can be conveniently optimized as required.
The method specifically comprises the following steps:
(1) Establishing a servo simulation model of a feeding system based on MATLAB;
(2) Determining a servo parameter optimization variable of a feed drive;
(3) Constructing a servo-optimized extensible dynamic performance evaluation function of the feeding system, selecting respective corresponding response characteristic index evaluation functions according to certain dynamic response performances to be comprehensively considered by the feeding system, combining actual optimization requirements of the feeding system, and substituting the functions into the extensible dynamic performance evaluation function to obtain a servo-optimized target evaluation function;
(4) Based on an intelligent optimization algorithm, the performance of the feeding system is optimized by servo parameter optimization in the direction of the constructed target evaluation function.
The optimization method of the invention is further explained by taking a certain machine tool feed driving system as an example and combining the attached drawings and an embodiment.
The method comprises the following steps: establishing servo simulation model of feeding system based on MATLAB
In general, the servo control of the feeding system usually adopts a classical cascade control method, which includes a current loop, a speed loop and a position loop. Transfer function methods are currently used to build feed system servo simulation models because the current loop is contained within the velocity loop and the velocity loop is contained within the position loop. Therefore, when a transfer function method is adopted to carry out simulation modeling on the servo control of the feeding system, firstly, a current loop is modeled, and then, a speed loop is further modeled, and finally, a position loop is modeled. The following modeling process is described in terms of this step.
The current loop mainly comprises a current regulator, a servo motor and the like, wherein the current controller is a PI controller and has the function of enabling the armature winding current of the servo motor to accurately track a current instruction in real time. The current loop controller transfer function may be expressed as K i (1+1/(T i s)), where K) is i For current loop gain, T i Is the current loop integration time. The transfer function model of armature circuit armature voltage and armature current of the servo motor can be obtained according to the control correlation theory as
In the formula e 0 Is a counter potential of K e ω(K e Is the back emf coefficient, ω is the output speed); i.e. i q Is armature current, u q Is armature voltage, L a Is inductance factor, R a Is the resistivity, s is the laplace operator. From the equation (1), it can be seen that the armature circuit of the servo motor has a time constant L a /R a The first-order inertia element.
The torque equation of the servo motor is
M T =K T i q (2)
In the formula K T As a coefficient of torque, M T Is the output torque.
The current loop is a follow-up system, current regulation is realized according to feedback errors of given current and actual current, and a current loop transfer function structure diagram can be obtained by integrating transfer functions and equations of all links, as shown in fig. 1, wherein i set The Δ i current feedback error is the current setpoint.
The speed loop also employs a PI controller, which is expressed as K in terms of a transfer function p (1+1/(T n s)), where K) is p For the gain of the speed loop, T n The velocity loop integration time. Speed regulationThe pitch ring is mainly composed of links such as a speed ring controller, a current closed loop and mechanical transmission.
The position loop adjuster adopts P adjustment and is expressed by a transfer function as K v s,K v Is the position loop gain.
A simulation model of the feed drive system is constructed based on MATLAB, a model schematic diagram of the simulation model is shown in fig. 2, wherein the mechanical transmission link modeling can be performed by simplifying the mechanical transmission link into an equivalent rotation rigid body, and can also be performed by a state space method, and the state space method is used for modeling in the embodiment. In FIG. 2, φ s To displace a set value, phi a Is the actual value of the displacement.
According to the servo control system and the servo motor model corresponding to the researched machine tool, the technical parameters adopted in the modeling process are as follows: position loop gain K v =25s -1 Velocity loop gain K p =27.3N · m · s/rad, velocity loop integration time T n =60ms, current loop gain K i =12.157V/A, current loop integration time T i 2ms, inductance factor L a =3.1mH, resistivity R a =0.075 Ω, back emf coefficient K e = 1.67V/(rad/s), torque coefficient K T =2.72N·m/A。
Step two: determining a servo parameter optimization variable of a feed drive;
generally, a cascaded PID control is adopted in a servo control system to optimize the gain and the integral time of a position loop, a speed loop and a current loop, but because the control parameters of the current loop can be obtained by a motor manufacturer and can be adjusted without a complex servo simulation model, three parameters of the gain and the integral time of the speed loop and the gain of the position loop are selected as optimization variables of servo parameters to optimize.
Step three: constructing a scalable dynamic performance evaluation function of the servo optimization of the feeding system;
defining response characteristic Evaluation Index functions (such as IAEI, ISE, ITAE and the like mentioned in the background) as System Performance Evaluation indexes (SPI), and constructing a Scalable Objective Evaluation Function (SOEF) of the servo dynamic response characteristics of the feeding System, wherein the Scaling Objective Evaluation Function (SOEF) comprises:
in formula (3), n is the number of system performance evaluation indexes SPI, w i Scalability factor, w, in the overall evaluation function for a single performance evaluation index i =0 or 1, as can be seen from formula (3) if w i The scalability coefficient is set to 0, and the corresponding performance evaluation index SPI does not have any influence on the overall evaluation function SOEF, i.e., the performance evaluation index is not considered in the optimization process of the response characteristic of the feeding system; if w i If the scalability coefficient is set to 1, the corresponding performance evaluation index SPI will have an influence on the overall evaluation function, i.e., the performance evaluation index is considered in the optimization process of the response characteristic of the feed system, and therefore, the scalability selection of the system optimization objective function can be achieved by selecting the scalability coefficient.
f i For system performance evaluation index SPI i When w is i If 1, can be measured by f i Value size to further control SPI i The degree of influence on the overall evaluation function. f. of i The value can be any positive number, and the value f can be a positive integer in general for convenient calculation i The larger the value is, the performance evaluation index SPI is represented i The larger the influence degree on the overall evaluation function is, so when the overall evaluation function or the dynamic performance optimization requirement emphasizes a certain performance evaluation index, the corresponding system performance evaluation index influence coefficient f i A larger value may be desirable.
The dynamic response performances of the feeding system, such as response speed, overshoot and transient oscillation, are comprehensively considered in the embodiment. Therefore, the ITSE criterion is selected to evaluate the response speed of the system and is used as the SPI 1 (ii) a Further, the formula (4) is selected as SPI 2 To evaluate overshoot and transient oscillation of the system:
in the formula (4), phi s To displace a set value, phi a Is the actual value of the displacement; a represents overshoot balance coefficient, b represents transient oscillation balance coefficient, and the values of a and b are arbitrary positive numbers p 0 Is in the form of p 1 t is a linear function or p of the time variable t 2 t 2 With respect to the time variable t, where p 1 Coefficient representing a linear function of time, p 2 Coefficients representing a quadratic time function; p is a radical of 0 、p 1 、p 2 The value is positive, and usually, the value is selected to be a positive integer for convenient calculation.
The overshoot trade-off coefficient a and the transient oscillation trade-off coefficient b have the following priorities: preference for positive numbers p 0 Then choose the shape as p 1 t is a time function of, and finally is selected to be p 2 t 2 A quadratic time function of (c). When the overshoot phenomenon is emphasized to the optimization requirement of the overshoot and transient oscillation response performance, the priority of the value of the corresponding overshoot weighing coefficient a is higher than that of the transient oscillation weighing coefficient b, and the values of a and b meet p 0 >p 1 >p 2 (ii) a For example, if a takes on a positive number p 0 The value of b is selected to be p 1 A time function of t or the form p 2 t 2 The second order time function of (a), correspondingly,
p 0 >p 1 or p 0 >p 2 (ii) a If a takes the form of p 1 t as a function of time, b is chosen to be p 2 t 2 Of a quadratic time function of, while p 1 >p 2 . On the contrary, when the transient oscillation phenomenon is emphasized more on the optimization requirement of the overshoot and transient oscillation response performance, the priority of the value of the corresponding transient oscillation weighing coefficient b is higher than that of the overshoot weighing coefficient a, and the values of a and b meet p 0 >p 1 >p 2 . After the value types (positive number, primary function or quadratic function) of a and b are determined according to optimization requirements and value priorities, corresponding p can be subjected to actual optimization conditions 0 、p 1 Or p 2 Is adjusted by the value of p 0 、p 1 Or p 2 The larger the value of (a) is, the larger the value of the corresponding coefficient a or b is, indicating that the overshoot corresponding to the coefficient a or the transient ringing corresponding to the coefficient b has a greater influence on the dynamic performance optimization result at the selected priority, and vice versa.
When the overshoot and transient oscillation response performance is optimized, and the optimization requirements of the overshoot phenomenon and the transient oscillation phenomenon are the same or similar in importance degree, the values of the corresponding overshoot weighing coefficient a and the transient oscillation weighing coefficient b are the same priority. For example, a and b may be shaped as p at the same time 2 t 2 The quadratic time function of (2) can also be taken simultaneously as p 1 A time function of t or a positive number p 0 。
According to the practical optimization requirements corresponding to the feeding system researched by the embodiment, namely, the dynamic response performances of the feeding system, such as response speed, overshoot and transient oscillation, are comprehensively considered, and w is defined
1 =w
2 =1 (which means that the response speed, overshoot, and transient oscillation of the system are considered at the same time), the response speed of the system is more emphasized when the dynamic performance is optimized, so f
1 =100,f
2 =1(f
1 Compared with f
2 Take a larger value to indicate that the response speed of the system is more emphasized when dynamic performance is optimized). In constructing SPI
2 When the overshoot and transient oscillation of the system are evaluated, the overshoot balance coefficient a is preferably selected to be the positive integer p because the overshoot phenomenon is more emphasized in the example
0 Selecting the transient oscillation trade-off coefficient b as p
2 t
2 The quadratic function of (2), in this example, p is chosen
0 =100,p
2 =25, so obtained SPI
2 =∫[100e
o +25t
2 e
s ]dt,
General purpose SPI 1 、SPI 2 And (3) substituting the formula (3) to obtain a target evaluation function of servo optimization:
step four: based on an intelligent optimization algorithm, the performance optimization of the feeding system is realized through servo parameter optimization and oriented to the constructed target evaluation function.
Based on an intelligent optimization algorithm, selecting a position loop gain K v Velocity loop gain K p Integral time T of velocity loop n For optimizing variables, the formula (5) is used as an optimization objective function to comprehensively evaluate the response speed, the overshoot and the transient oscillation of the system, the servo control parameters are optimized, the iterative process of objective function evolution is shown in figure 3, and finally the optimized parameters of the servo controller are K v =38s -1 ,K p =24Nm·s/rad,T n =19ms, the step response curves before and after optimization are as shown in fig. 4, and it can be seen from the graph that the dynamic response performance of the entire servo control system is good by using the optimized servo control parameters, the response speed of the system is fast (the rise time is 0.038 s) and has no overshoot, the adjustment time is 0.092s, the adjustment time is short, and the stability of the system is good, i.e., no longer time oscillation. These are consistent with preset dynamic performance optimization requirements.
On the basis of optimization of the embodiment, if the energy consumption in the system response process is further increased, the energy consumption is taken as an evaluation index SPI 3 Then, by definition w, with the aid of formula (3) 1 =w 2 =w 3 =1 and passes f according to actual requirements 3 Control SPI 3 The influence degree of the overall dynamic response evaluation function SOEF is substituted into the formula (3), so that the selected response characteristic evaluation index target function can be conveniently adjusted in real time and further integrated, and the performance of the servo system can be optimized as required.
It can be seen that the dynamic performance after the optimization by the method of the invention is obviously improved compared with that before the optimization. The optimization process is a continuously improved process, and in actual operation, values of parameters in the evaluation function can be further adjusted according to feedback of the optimization result, so that a better to optimal dynamic performance optimization result is obtained. For example according to f mentioned above i Value rule (f) i The performance evaluation index SP is represented by the larger the value isI i The larger the influence degree on the overall evaluation function is, when the overall evaluation function or the dynamic performance optimization requirement emphasizes a certain performance evaluation index, the corresponding system performance evaluation index influence coefficient f i May take a larger value), and p 0 、p 1 、p 2 Value rule (p) 0 、p 1 、p 2 The larger the value is, the larger the corresponding coefficient a or b is, which indicates that the overshoot phenomenon corresponding to the coefficient a or the transient oscillation phenomenon corresponding to the coefficient b has larger influence on the dynamic performance optimization result under the selected priority, and vice versa), and the p is readjusted 0 、p 1 、p 2 、f i And (5) optimizing the value.
For example: if it is desired to further increase the response speed of the system of this embodiment, f can be further increased 1 The value of (d); if the response speed and overshoot are found to meet the optimization requirement, but the transient oscillation phenomenon is further controlled, p can be further increased 2 The numerical value of (c).