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

Abstract

The invention relates to a control method for a large-scale electromechanical hybrid inertia system, which belongs to the field of motion control. After the large-scale electromechanical hybrid inertia system is initialized, set the expected output speed and inertia of the system, control the motor to adjust the output speed and inertia of the system; design the speed regulator and Kalman signal filter. The advantages are: taking into account the precise control of the system output inertia, meeting the control requirements of large-scale electromechanical hybrid inertia systems, improving the control accuracy and dynamic response performance of the control system, and greatly reducing the interference of space radiation, power supply, and signal transmission in the industrial field environment Factors interfere with the two, which improves the system feedback and control accuracy.

Description

A control method for a large electromechanical hybrid inertial system

technical field

The invention relates to a method in the field of motion control, in particular to a method for controlling the rotational speed and inertia of a large electromechanical hybrid inertia system based on a Kalman filter fuzzy PID algorithm.

Background technique

Large-scale electromechanical hybrid inertia systems are widely used in metallurgy, chemical industry, electric power, light industry and other fields. The purpose of loading is achieved mainly by controlling the motor drive and compensating the kinetic energy of the mechanical inertia. How to accurately and comprehensively control the motor speed and system output inertia is the technical key to controlling the loading of large electromechanical hybrid inertial systems. At present, in related fields, there are methods for the speed control of large-scale electromechanical hybrid inertia systems, but there are very few methods that can accurately control the output speed and inertia of the system at the same time. With the development of the industrial field, the application range of large-scale electromechanical hybrid inertia systems Increasingly, research on the method of precisely controlling its output speed and inertia is imminent.

In a large-scale electromechanical hybrid inertia control system, the adjustment of the motor speed is the technical premise of accurately controlling the output speed and inertia of the system. The PID controller is widely used in the speed control system due to its simple structure, strong robustness to model errors, and easy operation. For large-scale electromechanical hybrid inertia systems, the complexity of the controlled object continues to increase; and most large-scale electromechanical hybrid inertia systems need to change the mechanical inertia part at any time according to industrial needs, resulting in changes in system parameters and increasing the difficulty of precise control. Simple use of traditional PID control results in low control efficiency, poor system stability, and it is difficult to meet the precise requirements of target control.

In addition, because large-scale electromechanical hybrid inertial systems are mostly used in industrial environments with high interference, there are many random interferences, such as space radiation interference, power supply interference, and signal transmission interference, etc., which have a greater impact on the feedback control signal in the inertia control system , thus reducing the detection accuracy and control accuracy of the system. At present, the filtering methods for feedback signals in large-scale electromechanical hybrid inertial systems are not mature enough, and digital low-pass filtering techniques with low sensitivity are often used, which is difficult to meet the feedback signal filtering requirements of large-scale electromechanical hybrid inertial systems.

In summary, it is necessary to study a control method that can stably and accurately control large-scale electromechanical hybrid inertial systems.

Contents of the invention

The invention provides a control method for a large-scale electromechanical hybrid inertia system, aiming at precisely controlling the output rotational speed and inertia of the large-scale electromechanical hybrid inertia system.

The technical scheme adopted by the present invention is: comprise the following steps:

(1) After the large-scale electromechanical hybrid inertia system is initialized, set the expected output speed and inertia of the system;

(2) The electric inertia controller completes the calculation of the compensated electric inertia according to the set expected inertia combined with the rotational torque signal fed back by the torque sensor and processed by the Kalman signal filter, and outputs the correction gain value K. Output the given speed signal, the difference between the given speed signal and the speed feedback signal measured by the speed sensor and processed by the Kalman signal filter is used as the input signal of the speed regulator, and the output signal of the speed regulator is multiplied by the correction gain K to obtain the current The input signal of the regulator and the output signal of the current regulator output the current control signal of the motor through the thyristor rectifier device, so as to control the output speed and inertia of the motor regulation system;

(3) the speed regulator is designed based on the method of fuzzy controller online tuning PID parameters, to improve the dynamic response and steady-state characteristics of the speed control system;

(4) Take the sampling period as 0.01s to design the Kalman signal filter so that it can filter the speed feedback signal and torque signal feedback signal.

An embodiment of the present invention is: the specific design steps of the design speed regulator are as follows:

(a) Select rotational speed deviation e and deviation change rate e _c as input language variables, proportional gain coefficient K _p , integral gain coefficient K _i and differential gain coefficient K _d as output language variables, fuzzy level is seven, language variable values are {NB, NM, NS, ZO, PS, PM, PB} 7 fuzzy values, NB means negative big, NM means negative middle, NS means negative small, ZO means zero, PS means positive small, PM means positive middle, PB means Zhengda, the membership function is triangular membership function, according to the PID controller parameter self-tuning principle combined with expert experience to set the input and output language variable universe and fuzzy control rules;

(b) According to the fuzzy rules, the fuzzy controller output of each state is calculated off-line through the fuzzy logic reasoning method for all quantified combinations of all input language variables (speed deviation, deviation change rate), and finally the fuzzy control table is generated:

i) The control rule of K _p is:

ii) The control rule of K _i is:

iii) The control rule of K _p is:

Among them, the fuzzy reasoning machine adopts the Mamdani type reasoning system, and the defuzzifier adopts the center of gravity method. After the control table is obtained, the control table is stored in the computer, and a subroutine for searching the control table is compiled. In the actual control process, through the table lookup, Bring it into the formula calculation to get the adjusted K _p , K _i and K _d values.

An embodiment of the present invention is: the specific design method of the Kalman signal filter is as follows:

The state equation and observation equation of the control system model are discretized as a continuous system by using the zero-order holder method, and the sampling period is 0.01s. The discrete state equation and measurement equation are respectively:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Gamma;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\end{array}\right.<span\; class="notranslate">\; ,</span>$

In the formula: x(k)——state vector; y(k)——measurement 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 matrix;

According to the system state equation, the Kalman filter equation is obtained according to the Kalman filter theorem:

(a) One-step forecast estimation equation:

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

(b) The variance matrix of one-step forecast estimation error:

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

(d) Filter estimation equation:

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ]</span>$

(e) Variance matrix of filtering estimation error:

P _k|k ＝[IK _k C(k)]P _k|k-1

In the formula: Q(k), R(k) is the covariance matrix of random noise, I is the identity matrix, as long as the initial value of filtering is given P _0|0 , according to the measured value y(k) at time k, the state estimation value at time k can be obtained by recursive calculation

One embodiment of the present invention is that the large-scale electromechanical hybrid inertia system includes: an electric inertia controller, a speed regulator, a current regulator, a silicon controlled rectifier, a DC motor, a reducer, a mechanical inertia part and a Kalman signal filter device.

The advantages of the present invention are:

(1) The present invention is based on the double closed-loop structure of rotational speed control combined with the single closed-loop structure of torque control, and corrects the output inertia of the motor through the electric inertia controller. The output torque takes into account the precise control of the system output inertia, which meets the control requirements of the large-scale electromechanical hybrid inertia system.

(2) The rotating speed regulator of the present invention adopts the method of fuzzy algorithm on-line tuning PID parameter to design, overcomes the shortcomings such as strong volatility and poor adaptability of the traditional PID control method in the large-scale electromechanical hybrid inertia system, and solves the problem of large-scale electromechanical hybrid inertia The hysteresis and nonlinear problems of the system improve the control accuracy and dynamic response performance of the control system.

(3) The speed regulator designed by the present invention can adjust the PID parameters on-line through the fuzzy algorithm to adapt to the system parameter change caused by the change of the mechanical inertia part in the large-scale electromechanical hybrid inertia system, thereby improving the adaptability of the control system and expanding the range of different parameters. The scope of application of large-scale electromechanical hybrid inertia systems in the field.

(4) The present invention adopts the Kalman signal filter to filter the rotating speed and torque feedback signals of the large-scale electromechanical hybrid inertia system, which greatly reduces the interference of space radiation, power supply, signal transmission and other interference factors on the two in the industrial field environment. Improved system feedback and control accuracy.

Description of drawings

Fig. 1 is a schematic structural diagram of a large-scale electromechanical hybrid inertia control system of the present invention;

Fig. 2 is a schematic diagram of Kalman filter fuzzy PID control in the present invention;

Fig. 3 is a fuzzy regularizer input and output language variable membership degree function figure among the present invention;

Fig. 4 a is the fuzzy rule ΔK _p control chart in the specific embodiment of the present invention;

Fig. 4 b is the control figure of fuzzy rule ΔK _i in the specific embodiment of the present invention;

Fig. 4c is a control diagram of the fuzzy rule ΔK _d in the specific embodiment of the present invention.

Detailed ways

As shown in Figure 1, the large-scale electromechanical hybrid inertia control system based on Kalman filter fuzzy PID control of the present invention includes an electric inertia controller, a speed regulator, a current regulator, a silicon controlled rectifier, a DC motor, a speed reducer, a mechanical Inertia part, Kalman signal filter. In view of the impact of changing system performance on the control system, the fuzzy controller is used to adjust the PID parameters online to adjust the output speed of the motor speed control system; combined with the Kalman filter method to process the system speed and torque feedback signals to reduce industrial site environmental interference The impact on the system, precisely control the system output speed and inertia.

Based on the speed control double closed-loop combined with torque control single closed-loop structure: In the speed double closed-loop, the current loop is used as the inner loop, and the speed loop is used as the outer loop. Between the speed loop and the current loop, a torque correction parameter K is introduced to form a The speed and current closed loop of torque correction function. The difference y ^* -y between the actual output speed y of the system and the expected value y ^* of the speed is used as the input signal of the speed regulator. The output value of the speed regulator is multiplied by the torque correction parameter K to obtain the expected value i ^* of the current loop. The difference i ^* -i between the expected value of the current loop and the actual current i is used as the input signal of the current controller. The output signal of the current controller is used as the input signal v of the thyristor rectifier, and the output signal of the thyristor rectifier is the voltage signal v _d input to the DC motor. The input signal of the electric inertia controller includes the expected speed value y ^* , the expected inertia value I ^* , the installed mechanical inertia value I and the real-time output torque value T of the system, and the output is the torque correction parameter K, which is used to correct the output of the speed regulator Signal, the expected signal i ^* of the current loop has been obtained. Among them, the actual output speed signal and torque signal of the system are measured by the corresponding sensor and filtered through the Kalman signal filter, and the output current signal of the DC motor is measured by the corresponding sensor.

The control method of the large-scale electromechanical hybrid inertia system based on the Kalman filter fuzzy PID control comprises the following steps:

(1) After the large-scale electromechanical hybrid inertia system is initialized, set the expected output speed and inertia of the system;

(2) The electric inertia controller completes the calculation of the compensated electric inertia according to the set expected inertia combined with the rotational torque signal fed back by the torque sensor and processed by the Kalman signal filter, and outputs the correction gain value K. Output the given speed signal, the difference between the given speed signal and the speed feedback signal measured by the speed sensor and processed by the Kalman signal filter is used as the input signal of the speed regulator, and the output signal of the speed regulator is multiplied by the correction gain K to obtain the current The input signal of the regulator and the output signal of the current regulator output the current control signal of the motor through the thyristor rectifier device, so as to control the output speed and inertia of the motor regulation system;

(3) the speed regulator is designed based on the method of fuzzy controller online tuning PID parameters, to improve the dynamic response and steady-state characteristics of the speed control system;

(4) Take the sampling period as 0.01s to design the Kalman signal filter so that it can filter the speed feedback signal and torque signal feedback signal.

The specific design steps for designing the speed regulator described in this paper are as follows:

(a) Select rotational speed deviation e and deviation change rate e _c as input language variables, proportional gain coefficient K _p , integral gain coefficient K _i and differential gain coefficient K _d as output language variables, fuzzy level is seven, language variable values are {NB, NM, NS, ZO, PS, PM, PB} 7 fuzzy values, NB means negative big, NM means negative middle, NS means negative small, ZO means zero, PS means positive small, PM means positive middle, PB means Zhengda, the membership function is triangular membership function, according to the PID controller parameter self-tuning principle combined with expert experience to set the input and output language variable universe and fuzzy control rules;

(b) According to the fuzzy rules, the fuzzy controller output of each state is calculated off-line through the fuzzy logic reasoning method for all quantified combinations of all input language variables (speed deviation, deviation change rate), and finally the fuzzy control table is generated:

i) The control rule of K _p is:

ii) The control rule of K _i is:

iii) The control rule of K _p is:

Among them, the fuzzy reasoning machine adopts the Mamdani type reasoning system, and the defuzzifier adopts the center of gravity method. After the control table is obtained, the control table is stored in the computer, and a subroutine for searching the control table is compiled. In the actual control process, through the table lookup, Bring it into the formula calculation to get the adjusted K _p , K _i and K _d values.

The concrete design method of Kalman signal filter described in the present invention is as follows:

The state equation and observation equation of the control system model are discretized as a continuous system by using the zero-order holder method, and the sampling period is 0.01s. The discrete state equation and measurement equation are respectively:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Gamma;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\end{array}\right.<span\; class="notranslate">\; ,</span>$

In the formula: x(k)——state vector; y(k)——measurement 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 matrix;

According to the system state equation, the Kalman filter equation is obtained according to the Kalman filter theorem:

(a) One-step forecast estimation equation:

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

(b) The variance matrix of one-step forecast estimation error:

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

(d) Filter estimation equation:

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ]</span>$

(e) Variance matrix of filtering estimation error:

P _k|k ＝[IK _k C(k)]P _k|k-1

In the formula: Q(k), R(k) is the covariance matrix of random noise, I is the identity matrix, as long as the initial value of filtering is given P _0|0 , according to the measured value y(k) at time k, the state estimation value at time k can be obtained by recursive calculation

Claims (4)

1. A large-scale electromechanical hybrid inertial system control method is characterized in that comprising the following steps: (1) After the large-scale electromechanical hybrid inertia system is initialized, set the expected output speed and inertia of the system; (2) The electric inertia controller completes the calculation of the compensated electric inertia according to the set expected inertia combined with the rotational torque signal fed back by the torque sensor and processed by the Kalman signal filter, and outputs the correction gain value K. Output the given speed signal, the difference between the given speed signal and the speed feedback signal measured by the speed sensor and processed by the Kalman signal filter is used as the input signal of the speed regulator, and the output signal of the speed regulator is multiplied by the correction gain K to obtain the current The input signal of the regulator and the output signal of the current regulator output the current control signal of the motor through the thyristor rectifier device, so as to control the output speed and inertia of the motor regulation system; (3) the speed regulator is designed based on the method of fuzzy controller online tuning PID parameters, to improve the dynamic response and steady-state characteristics of the speed control system; (4) Take the sampling period as 0.01s to design the Kalman signal filter so that it can filter the speed feedback signal and torque signal feedback signal. 2. a kind of large electromechanical hybrid inertial system control method according to claim 1, is characterized in that the specific design steps of described design speed regulator are as follows: (a) Select rotational speed deviation e and deviation change rate e _c as input language variables, proportional gain coefficient K _p , integral gain coefficient K _i and differential gain coefficient K _d as output language variables, fuzzy level is seven, language variable values are {NB, NM, NS, ZO, PS, PM, PB} 7 fuzzy values, NB means negative big, NM means negative middle, NS means negative small, ZO means zero, PS means positive small, PM means positive middle, PB means Zhengda, the membership function is triangular membership function, according to the PID controller parameter self-tuning principle combined with expert experience to set the input and output language variable universe and fuzzy control rules; (b) According to the fuzzy rules, the fuzzy controller output of each state is calculated off-line through the fuzzy logic reasoning method for all quantified combinations of input language variables (speed deviation, deviation change rate), and the fuzzy control table is finally generated: i) The control rule of K _p is: ii) The control rule of K _i is: iii) The control rule of K _p is: Among them, the fuzzy reasoning machine adopts the Mamdani type reasoning system, and the defuzzifier adopts the center of gravity method. After the control table is obtained, the control table is stored in the computer, and a subroutine for searching the control table is compiled. In the actual control process, through the table lookup, Bring it into the formula calculation to get the adjusted K _p , K _i and K _d values. 3. a kind of large electromechanical hybrid inertial system control method according to claim 1, is characterized in that the specific design method of described Kalman signal filter is as follows: The state equation and observation equation of the control system model are discretized as a continuous system by using the zero-order holder method, and the sampling period is 0.01s. The discrete state equation and measurement equation are respectively: $\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Gamma;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>\end{array}\right.<span\; class="notranslate">\; ,</span>$ In the formula: x(k)——state vector; y(k)——measurement 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 matrix; According to the system state equation, the Kalman filter equation is obtained according to the Kalman filter theorem: (a) One-step forecast estimation equation: ${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$ (b) The variance matrix of one-step forecast estimation error: P _k|k-1 ＝A(k,k-1)P _k-1|k-1 A ^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-1 C ^T (k)[C(k)P _k|k-1 C ^T (k)+R(k)] ^-1 (d) Filter estimation equation: ${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ]</span>$ (e) Variance matrix of filtering estimation error: P _k|k ＝[IK _k C(k)]P _k|k-1 In the formula: Q(k), R(k) is the covariance matrix of random noise, I is the identity matrix, as long as the initial value of filtering is given P _0|0 , according to the measured value y(k) at time k, the state estimation value at time k can be obtained by recursive calculation 4. The control method of a large-scale electromechanical hybrid inertia system according to claim 1, wherein the large-scale electromechanical hybrid inertia system comprises: an electric inertia controller, a speed regulator, a current regulator, a silicon controlled rectifier, DC motor, reducer, mechanical inertia part and Kalman signal filter.