A method for suppressing thrust fluctuation of linear servo system

Abstract

The present invention relates to a method for suppressing thrust fluctuation of a linear servo system, and solves the problem of thrust fluctuation in the linear servo system. The method does not need to construct an offline mathematical model of ripple thrust, nor does it need to construct an online observer of ripple thrust. By deriving an analytical expression of a compensation current, a root mean square error value of speed fluctuation is constructed, and a mass estimation parameter in the analytical expression is adjusted according to the error through the output of a PI regulator, so that the error of the PI regulator tends to 0, thereby suppressing the ripple speed of the linear servo, and further suppressing the ripple thrust.

Description

Technical Field

[0001] The invention belongs to the field of automatic control, relates to linear servo system control, and in particular to a method for suppressing thrust fluctuation of a linear servo system. Background Art

[0002] The linear servo system has great advantages in driving high-speed and high-precision platforms of CNC machine tools. If the linear servo motor adopts an iron core structure, its cogging torque and end effect will cause the thrust of the servo motor to fluctuate when it runs within the travel range. This thrust fluctuation is called positioning force, which is related to the position of the mover. At the same time, the friction resistance of the linear servo system will change with the displacement due to the change of load resistance within the travel range and the change of the friction coefficient of the guide rail. These thrust fluctuations will cause vibration and noise in the feed motor, which will affect the system feed progress in severe cases, so it is necessary to suppress them.

[0003] The existing method of suppressing thrust fluctuation of linear servo, in addition to reducing thrust fluctuation by optimizing the tooth slot structure, the rotor core structure, the pole structure, the skewed slot or the skewed pole during motor design, can only use harmonic current injection for feedforward compensation of ripple thrust after the motor is manufactured. Harmonic current injection relies on the measurement of positioning force and FFT analysis to build an offline mathematical model or online observer of ripple thrust. The existing current compensation model is complex and insensitive to changes in the system quality of the linear servo system, which affects the fluctuation suppression effect. Summary of the invention

[0004] The purpose of the present invention is to solve the thrust fluctuation problem existing in the linear servo system mentioned in the background technology, and to provide a thrust fluctuation suppression method for the linear servo system, which does not require the construction of an offline mathematical model of ripple thrust, nor does it require the construction of an online observer for ripple thrust. By deriving an analytical expression for a compensation current, a speed fluctuation root mean square error value is constructed, and the parameters in the analytical expression are adjusted according to the error through the output of a PI regulator, so that the error of the PI regulator tends to 0, thereby achieving linear servo ripple speed suppression, and further achieving ripple thrust suppression.

[0005] The technical solution adopted by the present invention to solve the technical problem is: a method for suppressing thrust fluctuation of a linear servo system, characterized in that: when the linear servo system is running in a constant speed section, it is in a steady state with a given constant speed, at which time the ripple thrust f d Will result in a ripple speed v d , assuming that the thrust output by the motor without ripple in steady state is f e , the speed without ripple is v s , m is the system mass, which is composed of the linear motor mover, slider and load mass,

[0006] q-axis compensation current i qcThe calculation formula is as follows:

[0007]

[0008] f c The motor compensating thrust to eliminate the ripple force, f c +f d =0, k t is the thrust constant of the linear servo motor, which represents the ratio between the thrust of the linear servo motor and the q-axis current, v is the real-time speed, and x is the displacement value;

[0009] The calculation of the discretized compensation current is as follows:

[0010]

[0011] After the workpiece is replaced, the system mass m will also change; therefore, the compensation current is calculated based on the above formula. Since the system mass m is unknown, the estimated value of the mass is obtained through online adjustment. Formula 1 is rewritten as Formula 2:

[0012]

[0013] To obtain Define the comprehensive error of ripple speed as v d_rms (i), i is an integer greater than or equal to 0, that is, the error root mean square is obtained by rolling method, and the velocity error root mean square v at the current moment is obtained based on N sampling points within a pole pitch range of the linear servo motor, that is, within the range of [x-τ, x]. d_rms (i)

[0014]

[0015] Assume the control period of the speed loop is T v , then N is calculated by the following formula, where int means rounding the result, and v at constant speed s and T v are all constants;

[0016]

[0017] By defining the speed fluctuation root mean square error e(i):

[0018] e(i)=v d_rms (i)-v d_rms (i-1)

[0019] The quality estimation algorithm is: take e(i) as the error input of the PI controller, and As the output of the PI controller, when the error e(i)>0, it means If it is too small, then increase To reduce the speed ripple; when the error e(i) < 0, it means If too large, reduce To reduce the speed ripple; when the error e(i) = 0, it means It is already the optimal value;

[0020] Based on real-time online quality estimates Combining the discretized compensation current formula 2, the q-axis compensation current i can be obtained by calculation: qc (i), the q-axis compensation current i qc (i) As feedforward, it is superimposed on the q-axis current command i of the linear servo motor. qref (i) can get the compensated q-axis current command i qref_new (i)=i qref (i)+i qc (i) The ripple thrust can be compensated and offset by current closed-loop control, thereby achieving ripple thrust suppression.

[0021] The present invention constructs a speed fluctuation root mean square error value, and adjusts the mass estimation parameter in the analytical expression through the output of a PI regulator according to the error, thereby making the error of the PI regulator tend to 0, achieving linear servo ripple speed suppression, and further achieving ripple thrust suppression. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The present invention will be further described below in conjunction with the accompanying drawings.

[0023] Figure 1 It is a block diagram of the linear servo system quality estimation process of the present invention.

[0024] Figure 2 It is a block diagram of the thrust fluctuation suppression process of the linear servo system of the present invention. DETAILED DESCRIPTION

[0025] The present invention will be further described below through specific embodiments in conjunction with the accompanying drawings.

[0026] Embodiment, a maximum power capture control method for a direct-drive wave energy power generation device, such as Figure 1-2 As shown. According to the linear servo in the constant speed section, it is in a steady state with a constant speed. At this time, the ripple thrust f d Will result in a ripple speed v d , assuming that the thrust output by the motor without ripple in steady state is f e , the speed without ripple is v s , m is the system mass, which consists of the linear motor rotor, slider and load mass, and the load resistance is f L , then the motion equation of the linear servo system is as follows:

[0027]

[0028] The motor thrust consists of the thrust without ripple and the ripple thrust, then:

[0029]

[0030] In steady state:

[0031]

[0032] and then:

[0033]

[0034] If the motor thrust can be compensated to offset the influence of ripple thrust, the speed fluctuation component v d can be eliminated. Therefore, the compensating thrust f c Should meet:

[0035] f c +f d =0

[0036] Therefore:

[0037]

[0038] The ripple thrust of the actual linear servo motor changes with the displacement x of the motor rotor, so the relationship of the compensation thrust can be expressed in one step as:

[0039]

[0040] When the linear servo motor is running at a constant speed, it can be further expressed as:

[0041]

[0042] The thrust of the linear servo motor is proportional to the q-axis current, so the q-axis compensation current i qc The calculation formula is as follows:

[0043]

[0044] where k t is the thrust constant of the linear servo motor, which represents the ratio between the thrust of the linear servo motor and the q-axis current, and v is the real-time speed.

[0045] The calculation of the discretized compensation current adopts formula 1:

[0046]

[0047] The mass m of the actual linear servo system includes the mass of the linear motor rotor and the mass of the load workpiece. When the workpiece is replaced, m will also change. Therefore, the compensation current is calculated based on the above formula. Since the mass m of the system is unknown, it is necessary to obtain an estimated value of the mass through online adjustment. Formula 1 is rewritten as Formula 2:

[0048]

[0049] To obtain Define the comprehensive error of ripple speed as v d_rms (i), i is an integer greater than or equal to 0, that is, the error root mean square is obtained by rolling method, and the velocity error root mean square v at the current moment is obtained based on N sampling points within a pole pitch range of the linear servo motor, that is, within the range of [x-τ, x]. d_rms (i)

[0050]

[0051] Assume the control period of the speed loop is T v , then N is calculated by the following formula, where int means rounding the result, and v at constant speed s and T v are all constants.

[0052]

[0053] By defining the speed fluctuation root mean square error e(i):

[0054] e(i)=v d_rms (i)-v d_rms (i-1)

[0055] The quality estimation algorithm is: take e(i) as the error input of the PI controller, and As the output of the PI controller, when the error e(i)>0, it means Too small, need to be enlarged To reduce the speed ripple; when the error e(i) < 0, it means Too big, need to be reduced To reduce the speed ripple; when the error e(i) = 0, it means It is already the optimal value.

[0056] Based on real-time online quality estimates Combining the discretized compensation current formula 2, the q-axis compensation current i can be obtained by calculation: qc (i).

[0057] like Figure 2 As shown, the q-axis compensation current i qc(i) As feedforward, it is superimposed on the q-axis current command i of the linear servo motor. qref (i) can get the compensated q-axis current command i qref_new (i) Through current closed-loop control, the ripple thrust can be compensated and offset, thus achieving ripple thrust suppression. qref_new (i) The calculation formula is as follows: qref_new (i)=i qref (i)+i qc (i).

Claims

1. A method for suppressing thrust fluctuation of a linear servo system, characterized in that: When the linear servo system is running in the constant speed section, it is in a steady state with a given constant speed. At this time, the ripple thrust f d Will result in a ripple speed v d , assuming that the thrust output by the motor without ripple in steady state is f e , the speed without ripple is v s , m is the system mass, which is composed of the linear motor mover, slider and load mass, q-axis compensation current i qc The calculation formula is as follows: f c The motor compensating thrust to eliminate the ripple force, f c +f d =0, k t is the thrust constant of the linear servo motor, which represents the ratio between the thrust of the linear servo motor and the q-axis current, v is the real-time speed, and x is the displacement value; The calculation of the discretized compensation current is as follows: After the workpiece is replaced, the system mass m will also change; therefore, the compensation current is calculated based on the above formula. Since the system mass m is unknown, the estimated value of the mass is obtained through online adjustment. Formula 1 is rewritten as Formula 2: To obtain Define the comprehensive error of ripple speed as v d_rms (i), i is an integer greater than or equal to 0, that is, the error root mean square is obtained by rolling method, and the velocity error root mean square v at the current moment is obtained based on N sampling points within a pole pitch τ of the linear servo motor, that is, within the range of [x-τ,x]. d_rms (i) Assume the control period of the speed loop is T v , then N is calculated by the following formula, where int means rounding the result, and v at constant speed s and T v are all constants; By defining the speed fluctuation root mean square error e(i): e(i)=v d_rms (i)-v d_rms (i-1) The quality estimation algorithm is: take e(i) as the error input of the PI controller, and As the output of the PI controller, when the error e(i)>0, it means If it is too small, then increase To reduce the speed ripple; when the error e(i) < 0, it means If too large, reduce To reduce the speed ripple; when the error e(i) = 0, it means It is already the optimal value; Based on real-time online quality estimates Combining the discretized compensation current formula 2, the q-axis compensation current i can be obtained by calculation: qc (i), the q-axis compensation current i qc (i) As feedforward, it is superimposed on the q-axis current command i of the linear servo motor. qref (i) can get the compensated q-axis current command i qref_new (i)=i qref (i)+i qc (i) The ripple thrust can be compensated and offset by current closed-loop control, thereby achieving ripple thrust suppression.

Images