CN104238359A - Control method of large electromechanical mixed inertia system - Google Patents

Abstract

The invention relates to a control method of a large electromechanical mixed inertia system, and belongs to the field of motion control. After the large electromechanical mixed inertia system is initialized, the expected output rotating speed and the output inertia of the system are set, and a motor is controlled to adjust the output rotating speed and the output inertia of the system; a rotating speed adjustor and a Kalman signal filter are designed. The control method has the advantages that the accurate control over the output inertia of the system is considered, the control requirement for the large electromechanical mixed inertia system is met, the control accuracy and the dynamic response performance of the control system are improved, the interference with the control accuracy and the dynamic response performance by interference factors such as space radiation, a power source, and signal transmission in the industrial site environment is greatly reduced, and the feedback and control accuracy of the system is improved.

Description

A kind of large-scale electromechanics mixing inertia system control method

Technical field

The present invention relates to a kind of method of motion control field, be specifically related to a kind of method rotating speed of large-scale electromechanics mixing inertia system, inertia controlled based on Kalman filtering fuzzy PID algorithm.

Background technology

Large-scale electromechanics mixing inertia system is widely used in the fields such as metallurgy, chemical industry, electric power, light industry, driving and compensating mechanical inertia part kinetic energy, reaching loading object mainly through controlling motor.How exactly, control motor speed and system with taking into account to export inertia be control the key problem in technology that large-scale electromechanics mixing inertia system carries out loading.At present in the related art, there is the method controlled for large-scale electromechanics mixing inertia system rotating speed, but the method for accuracy-control system output speed and inertia is few while of energy, along with the development of industrial circle, large-scale electromechanics mixing inertia system range of application expanding day, extremely urgent to the technique study accurately controlling its output speed and inertia.

In large-scale electromechanics mixing inertia control system, the adjustment of motor speed is the technology prerequisite of accuracy-control system output speed and inertia.PID controller due to its structure simple, to model error, there is strong robustness and be easy to the advantages such as operation, be widely used in revolution speed control system, but along with industrial development, especially for the large-scale electromechanics mixing inertia system with large time delay, time variation and non-linear behavior, controlled device complexity is constantly deepened; And most large-scale electromechanics mixing inertia system will change mechanical inertia part at any time according to industrial requirement, causes systematic parameter to change, adds the difficulty accurately controlled.The traditional PID of simple employing controls to cause control efficiency low, and system stability is poor, is difficult to the precision requirement meeting target control.

In addition, because large-scale electromechanics mixing inertia system is applied in the larger industrial environment of interference more, there is many random disturbance, as space radiation interference, power supply disturbance and Signal transmissions interference etc., larger impact is produced on feedback control signal in inertia control system, thus reduces accuracy of detection and the control accuracy of system.Ripe not enough for the filtering method of feedback signal in large-scale electromechanics mixing inertia system at present, the digital low-pass filtering technology adopting sensitivity lower, is difficult to meet large-scale electromechanics mixing inertia system feedback signal filtering requirements more.

To sum up, it is very necessary for studying the control method a kind ofly stablizing, accurately to control large-scale electromechanics mixing inertia system.

Summary of the invention

The invention provides a kind of large-scale electromechanics mixing inertia system control method, object is accurately to control large-scale electromechanics mixing inertia system output speed and inertia.

The present invention takes technical scheme to be: comprise the following steps:

(1) after large-scale electromechanics mixing inertia system initialization, initialization system desired output rotating speed and inertia;

(2) electrical inertia controller completes the calculating compensating electrical inertia in conjunction with moment sensor feedback and through the rotating torque signal of Kalman's traffic filter process according to the expectation inertia of setting, and export correction gain value K, electrical inertia controller exports given tach signal according to expectation rotating speed, given rotating speed signal and speed probe measure and the difference of the speed feedback signal obtained through the process of Kalman's traffic filter as the input signal of speed regulator, speed regulator output signal is multiplied by modified gain K, obtain the input signal of current regulator, the output signal of current regulator is through the current controling signal of silicon controlled rectifier output motor, thus control electric machine regulating system output speed and inertia,

(3) described speed regulator designs based on the method for fuzzy controller on-line tuning pid parameter, to improve dynamic response and the steady-state characteristic of revolution speed control system;

(4) getting the sampling period is that 0.01s designs Kalman's traffic filter, can carry out filtering process to speed feedback signal and torque letter feedback number.

One embodiment of the present invention is: the specific design step of described design speed regulator is as follows:

A () chooses rotating speed deviation e and deviation variation rate e _cas input language variable, proportional gain factor K _p, integration gain factor K _iwith differential gain COEFFICIENT K _das output language variable, fuzzy class is seven grades, and linguistic variable value gets { NB, NM, NS, ZO, PS, PM, PB}7 fuzzy value, NB represents negative large, NM represent negative in, NS represents negative little, ZO represents that zero, PS represents just little, and PM represents center, PB represents honest, and subordinate function all selects Triangleshape grade of membership function, according to PID controller parameter Self-tuning System principle in conjunction with expertise setting input, output language variable field and fuzzy control rule;

B () is according to fuzzy rule, to all input language variablees (rotating speed deviation, deviation variation rate) quantize after various be combined through fuzzy logic inference method calculated off-line go out each state fuzzy controller export, finally generate fuzzy control table:

I) K _pcontrol law be:

Ii) K _icontrol law be:

Iii) K _pcontrol law be:

Wherein indistinct logic computer adopts Mamdani type inference system, and defuzzifier adopts gravity model appoach, after trying to achieve control table, control table is stored in a computer, and work out the subroutine that is searched control table, by tabling look-up in working control process, bringing formulae discovery into and can obtain the K after adjusting _p, K _iand K _dvalue.

One embodiment of the present invention is: described Kalman's traffic filter specific design method is as follows:

Adopt zero-order holder method that the state equation of control system model and observation equation are carried out discrete model construction, get sampling period 0.01s, its discrete state equation and measurement equation are respectively:

$\left\{\begin{array}{c}x(k+1)=A(k+1,k)x\left(k\right)+B(k+1,k)u\left(k\right)+\mathrm{\Γ}(k+1,k)w\left(k\right)\\ y\left(k\right)=C\left(k\right)x\left(k\right)+v\left(k\right)\end{array}\right.,$

In formula: x (k)---state vector; Y (k)---measure vector; U (k)---input vector (control signal); W (k)---control interference noise vector; V (k)---measurement noise vector; A (k+1, k)---system matrix; B (k+1, k)---input matrix; C (k)---output matrix; Γ (k+1, k)---constant matrices;

For system state equation, obtain Kalman filter equation according to Kalman filtering theorem:

(a) one-step prediction estimate equation:

${\hat{x}}_{k|k-1}=A(k,k-1){\hat{x}}_{k-1|k-1}+B(k,k-1)u(k-1)$

(b) one-step prediction variance of estimaion error battle array:

P _k|k-1＝A(k,k-1)P _k-1|k-1A ^T(k,k-1)+Γ(k,k-1)Q(k-1)Γ ^T(k,k-1)

(c) filter gain equation:

K _k＝P _k|k-1C ^T(k)[C(k)P _k|k-1C ^T(k)+R(k)] ^-1

(d) filtering estimate equation:

${\hat{x}}_{k|k}={\hat{x}}_{k-1|k-1}+K_k[y_k-C\left(k\right){\hat{x}}_{k|k-1}]$

(e) filtering variance of estimaion error battle array:

P _k|k＝[I-K _kC(k)]P _k|k-1

In formula: the covariance matrix that Q (k), R (k) are random noise, I is unit matrix, as long as given filtering initial value p _0|0, just can be obtained the state estimation in k moment by recurrence calculation according to measuring value y (k) in k moment

One embodiment of the present invention is, described large-scale electromechanics mixing inertia system comprises: electrical inertia controller, speed regulator, current regulator, silicon controlled rectifier, direct current generator, speed reduction unit, mechanical inertia part and Kalman's traffic filter.

Advantage of the present invention is:

(1) the present invention is based on rotating speed control two close cycles combined torque and control single closed loop configuration, export inertia by electrical inertia controller to motor to revise, while the rotating speed of large-scale electromechanics mixing inertia system is controlled, take into account by regulating system transmission shaft Driving Torque the accurate control that system exports inertia, met large-scale electromechanics mixing inertia Systematical control demand.

(2) speed regulator device of the present invention adopts the method for fuzzy algorithm on-line tuning pid parameter to design, the shortcomings such as the undulatory property overcoming traditional PID control method in large-scale electromechanics mixing inertia system is strong, bad adaptability, solve hysteresis quality and the nonlinear problem of large-scale electromechanics mixing inertia system, improve control accuracy and the dynamic response performance of control system.

(3) speed regulator of the present invention's design can be changed with the systematic parameter adapting to cause due to the change of mechanical inertia part in large-scale electromechanics mixing inertia system by fuzzy algorithm on-line tuning pid parameter, thus improve the adaptability of control system, expand the different field medium-and-large-sized electromechanics mixing inertia system scope of application.

(4) the present invention adopts Kalman's traffic filter to carry out filtering process to large-scale electromechanics mixing inertia system rotating speed, torque feedback signal, greatly reduce the disturbing factors such as space radiation in industrial environment, power supply, Signal transmissions to the interference of the two, improve system feedback and control accuracy.

Accompanying drawing explanation

Fig. 1 is the present invention's large-scale electromechanics mixing inertia Control system architecture schematic diagram;

Fig. 2 is Kalman filtering fuzzy-adaptation PID control schematic diagram in the present invention;

Fig. 3 is fuzzy rule device input and output linguistic variable membership function figure in the present invention;

Fig. 4 a is fuzzy rule Δ K in specific embodiments of the invention _pcontrol chart;

Fig. 4 b is fuzzy rule Δ K in specific embodiments of the invention _icontrol chart;

Fig. 4 c is fuzzy rule Δ K in specific embodiments of the invention _dcontrol chart.

Embodiment

As shown in Figure 1, the large-scale electromechanics mixing inertia control system that the present invention is based on Kalman filtering fuzzy-adaptation PID control comprises electrical inertia controller, speed regulator, current regulator, silicon controlled rectifier, direct current generator, speed reduction unit, mechanical inertia part, Kalman's traffic filter.For the system performance changed on the impact of control system, the method for fuzzy controller on-line tuning pid parameter is adopted to regulate Motor Rotating Speed Control System output speed; In conjunction with kalman filter method, system rotating speed, torque feedback signal are processed, reduce the impact of industrial environment interference on system, accuracy-control system output speed and inertia.

Control two close cycles combined torque based on rotating speed and control single closed loop configuration: in rotating speed two close cycles, current return as inner looping, speed loop as external loop, between speed loop and current return, introduce torque modification parameter K, form speed, the current closed-loop with torque modification function.The actual output speed y of system and the expectation value y of rotating speed ^*difference y ^*-y is as the input signal of speed regulator.The output valve of speed regulator is multiplied by torque modification parameter K, obtains the expectation value i of current return ^*.The expectation value of current return and the difference i of actual current i ^*-i is as the input signal of current controller.The output signal of current controller is as the input signal v of silicon controlled rectifier, and the output signal of silicon controlled rectifier is for inputing to the voltage signal v of direct current generator _d.The input signal of electrical inertia controller comprises rotating speed expectation value y ^*, inertia expectation value I ^*, the mechanical inertia value I that installs and the real-time Driving Torque value T of system, export as torque modification parameter K, K are used for revising the output signal of speed regulator, obtained the wanted signal i of current return ^*.Wherein, the actual output speed signal of system and dtc signal are by the sensor measurement responded and carry out filtering process acquisition by Kalman's traffic filter, and the output current signal of direct current generator is recorded by corresponding sensor.

The described large-scale electromechanics mixing inertia system control method based on Kalman filtering fuzzy-adaptation PID control, comprises the following steps:

(1) after large-scale electromechanics mixing inertia system initialization, initialization system desired output rotating speed and inertia;

(2) electrical inertia controller completes the calculating compensating electrical inertia in conjunction with moment sensor feedback and through the rotating torque signal of Kalman's traffic filter process according to the expectation inertia of setting, and export correction gain value K, electrical inertia controller exports given tach signal according to expectation rotating speed, given rotating speed signal and speed probe measure and the difference of the speed feedback signal obtained through the process of Kalman's traffic filter as the input signal of speed regulator, speed regulator output signal is multiplied by modified gain K, obtain the input signal of current regulator, the output signal of current regulator is through the current controling signal of silicon controlled rectifier output motor, thus control electric machine regulating system output speed and inertia,

(3) described speed regulator designs based on the method for fuzzy controller on-line tuning pid parameter, to improve dynamic response and the steady-state characteristic of revolution speed control system;

(4) getting the sampling period is that 0.01s designs Kalman's traffic filter, can carry out filtering process to speed feedback signal and torque letter feedback number.

The specific design step of this described design speed regulator is as follows:

A () chooses rotating speed deviation e and deviation variation rate e _cas input language variable, proportional gain factor K _p, integration gain factor K _iwith differential gain COEFFICIENT K _das output language variable, fuzzy class is seven grades, and linguistic variable value gets { NB, NM, NS, ZO, PS, PM, PB}7 fuzzy value, NB represents negative large, NM represent negative in, NS represents negative little, ZO represents that zero, PS represents just little, and PM represents center, PB represents honest, and subordinate function all selects Triangleshape grade of membership function, according to PID controller parameter Self-tuning System principle in conjunction with expertise setting input, output language variable field and fuzzy control rule;

B () is according to fuzzy rule, to all input language variablees (rotating speed deviation, deviation variation rate) quantize after various be combined through fuzzy logic inference method calculated off-line go out each state fuzzy controller export, finally generate fuzzy control table:

I) K _pcontrol law be:

Ii) K _icontrol law be:

Iii) K _pcontrol law be:

Wherein indistinct logic computer adopts Mamdani type inference system, and defuzzifier adopts gravity model appoach, after trying to achieve control table, control table is stored in a computer, and work out the subroutine that is searched control table, by tabling look-up in working control process, bringing formulae discovery into and can obtain the K after adjusting _p, K _iand K _dvalue.

Kalman's traffic filter specific design method of the present invention is as follows:

Adopt zero-order holder method that the state equation of control system model and observation equation are carried out discrete model construction, get sampling period 0.01s, its discrete state equation and measurement equation are respectively:

$\left\{\begin{array}{c}x(k+1)=A(k+1,k)x\left(k\right)+B(k+1,k)u\left(k\right)+\mathrm{\Γ}(k+1,k)w\left(k\right)\\ y\left(k\right)=C\left(k\right)x\left(k\right)+v\left(k\right)\end{array}\right.,$

In formula: x (k)---state vector; Y (k)---measure vector; U (k)---input vector (control signal); W (k)---control interference noise vector; V (k)---measurement noise vector; A (k+1, k)---system matrix; B (k+1, k)---input matrix; C (k)---output matrix; Γ (k+1, k)---constant matrices;

For system state equation, obtain Kalman filter equation according to Kalman filtering theorem:

(a) one-step prediction estimate equation:

${\hat{x}}_{k|k-1}=A(k,k-1){\hat{x}}_{k-1|k-1}+B(k,k-1)u(k-1)$

(b) one-step prediction variance of estimaion error battle array:

P _k|k-1＝A(k,k-1)P _k-1|k-1A ^T(k,k-1)+Γ(k,k-1)Q(k-1)Γ ^T(k,k-1)

(c) filter gain equation:

K _k＝P _k|k-1C ^T(k)[C(k)P _k|k-1C ^T(k)+R(k)] ^-1

(d) filtering estimate equation:

${\hat{x}}_{k|k}={\hat{x}}_{k-1|k-1}+K_k[y_k-C\left(k\right){\hat{x}}_{k|k-1}]$

(e) filtering variance of estimaion error battle array:

P _k|k＝[I-K _kC(k)]P _k|k-1

In formula: the covariance matrix that Q (k), R (k) are random noise, I is unit matrix, as long as given filtering initial value p _0|0, just can be obtained the state estimation in k moment by recurrence calculation according to measuring value y (k) in k moment

Claims (4)

1. a large-scale electromechanics mixing inertia system control method, is characterized in that comprising the following steps:

(1) after large-scale electromechanics mixing inertia system initialization, initialization system desired output rotating speed and inertia;

(2) electrical inertia controller completes the calculating compensating electrical inertia in conjunction with moment sensor feedback and through the rotating torque signal of Kalman's traffic filter process according to the expectation inertia of setting, and export correction gain value K, electrical inertia controller exports given tach signal according to expectation rotating speed, given rotating speed signal and speed probe measure and the difference of the speed feedback signal obtained through the process of Kalman's traffic filter as the input signal of speed regulator, speed regulator output signal is multiplied by modified gain K, obtain the input signal of current regulator, the output signal of current regulator is through the current controling signal of silicon controlled rectifier output motor, thus control electric machine regulating system output speed and inertia,

(3) described speed regulator designs based on the method for fuzzy controller on-line tuning pid parameter, to improve dynamic response and the steady-state characteristic of revolution speed control system;

(4) getting the sampling period is that 0.01s designs Kalman's traffic filter, can carry out filtering process to speed feedback signal and torque letter feedback number.

2. one according to claim 1 large-scale electromechanics mixing inertia system control method, is characterized in that the specific design step of described design speed regulator is as follows:

A () chooses rotating speed deviation e and deviation variation rate e _cas input language variable, proportional gain factor K _p, integration gain factor K _iwith differential gain COEFFICIENT K _das output language variable, fuzzy class is seven grades, and linguistic variable value gets { NB, NM, NS, ZO, PS, PM, PB}7 fuzzy value, NB represents negative large, NM represent negative in, NS represents negative little, ZO represents that zero, PS represents just little, and PM represents center, PB represents honest, and subordinate function all selects Triangleshape grade of membership function, according to PID controller parameter Self-tuning System principle in conjunction with expertise setting input, output language variable field and fuzzy control rule;

B () is according to fuzzy rule, to all input language variablees (rotating speed deviation, deviation variation rate) quantize after various be combined through fuzzy logic inference method calculated off-line go out each state fuzzy controller export, finally generate fuzzy control table:

I) K _pcontrol law be:

Ii) K _icontrol law be:

Iii) K _pcontrol law be:

Wherein indistinct logic computer adopts Mamdani type inference system, and defuzzifier adopts gravity model appoach, after trying to achieve control table, control table is stored in a computer, and work out the subroutine that is searched control table, by tabling look-up in working control process, bringing formulae discovery into and can obtain the K after adjusting _p, K _iand K _dvalue.

3. one according to claim 1 large-scale electromechanics mixing inertia system control method, is characterized in that described Kalman's traffic filter specific design method is as follows:

Adopt zero-order holder method that the state equation of control system model and observation equation are carried out discrete model construction, get sampling period 0.01s, its discrete state equation and measurement equation are respectively:

$\left\{\begin{array}{c}x(k+1)=A(k+1,k)x\left(k\right)+B(k+1,k)u\left(k\right)+\mathrm{\Γ}(k+1,k)w\left(k\right)\\ y\left(k\right)=C\left(k\right)x\left(k\right)+v\left(k\right)\end{array}\right.,$

In formula: x (k)---state vector; Y (k)---measure vector; U (k)---input vector (control signal); W (k)---control interference noise vector; V (k)---measurement noise vector; A (k+1, k)---system matrix; B (k+1, k)---input matrix; C (k)---output matrix; Γ (k+1, k)---constant matrices;

For system state equation, obtain Kalman filter equation according to Kalman filtering theorem:

(a) one-step prediction estimate equation:

${\hat{x}}_{k|k-1}=A(k,k-1){\hat{x}}_{k-1|k-1}+B(k,k-1)u(k-1)$

(b) one-step prediction variance of estimaion error battle array:

P _k|k-1＝A(k,k-1)P _k-1|k-1A ^T(k,k-1)+Γ(k,k-1)Q(k-1)Γ ^T(k,k-1)

(c) filter gain equation:

K _k＝P _k|k-1C ^T(k)[C(k)P _k|k-1C ^T(k)+R(k)] ^-1

(d) filtering estimate equation:

${\hat{x}}_{k|k}={\hat{x}}_{k-1|k-1}+K_k[y_k-C\left(k\right){\hat{x}}_{k|k-1}]$

(e) filtering variance of estimaion error battle array:

P _k|k＝[I-K _kC(k)]P _k|k-1

In formula: the covariance matrix that Q (k), R (k) are random noise, I is unit matrix, as long as given filtering initial value p _0|0, just can be obtained the state estimation in k moment by recurrence calculation according to measuring value y (k) in k moment

4. one according to claim 1 large-scale electromechanics mixing inertia system control method, is characterized in that described large-scale electromechanics mixing inertia system comprises: electrical inertia controller, speed regulator, current regulator, silicon controlled rectifier, direct current generator, speed reduction unit, mechanical inertia part and Kalman's traffic filter.