CN-122007308-A - Track correction and control method for multi-axis electro-hydraulic servo system of radial forging machine

CN122007308ACN 122007308 ACN122007308 ACN 122007308ACN-122007308-A

Abstract

The invention discloses a track correction and control method of a multi-axis electrohydraulic servo system of a radial forging machine, and provides a cooperative control scheme for fusing an extended state observer and online iterative learning, aiming at the complex control problem of coexistence of random high-frequency forging impact and workpiece axial plastic extension accumulated deviation in the radial forging process. The method comprises the steps of establishing and discretizing a system model, carrying out real-time estimation and feedforward compensation on total disturbance including random impact by using an extended state observer, carrying out multi-step state prediction based on the discrete model, carrying out iterative correction on prediction errors caused by slow-varying factors such as workpiece extension by adopting an online iterative learning algorithm, and generating a track correction amount. Finally, the disturbance compensation and the track correction amount are jointly acted on the controller, so that synchronous inhibition and compensation of high-frequency random disturbance and low-frequency trend deviation are realized, and the track tracking precision and robustness of the multi-axis system under strong impact and nonlinear working conditions are remarkably improved.

Inventors

PI YANGJUN
CHEN YACHUN
Tie Yuanhao
HU YI
CHENG MIN
WANG XIANSHENG
JIANG LIMING
HU YAO

Assignees

重庆大学

Dates

Publication Date: 20260512
Application Date: 20260210

Claims (10)

1. A track correction and control method of a multi-axis electro-hydraulic servo system of a radial forging machine is characterized by comprising the following steps: S1, acquiring real operation data of the multi-axis electro-hydraulic servo system under a forging working condition, wherein the real operation data comprise rod cavity pressure, rodless cavity pressure and piston rod displacement of a hydraulic cylinder; s2, based on the real operation data acquired in the step S1, a nonlinear state space equation model of the valve control hydraulic cylinder is established, and the model is simplified into a standard nonlinear state space equation model; S3, determining a steady-state working point of the system based on the real operation data acquired in the step S1, and carrying out linearization processing on the standard nonlinear state space equation model obtained in the step S2 at the steady-state working point to obtain a linearization state space equation; s4, deducing an open loop transfer function of the system based on the linearized state space equation obtained in the step S3, and introducing a PID controller to construct a closed loop control system to obtain the closed loop transfer function of the system; S5, converting the closed loop transfer function obtained in the step S4 into a discrete state space equation form; s6, planning an ideal tracking track from point to point for each axis of the multi-axis electro-hydraulic servo system; S7, predicting the system state of a plurality of sampling moments in the future by adopting a recursion method based on the discrete state space equation obtained in the step S5 and the ideal tracking track planned in the step S6; S8, calculating a track correction quantity at a future control moment by using an online iterative learning algorithm according to the error between the system state predicted in the step S7 and the ideal tracking track at the corresponding moment, and adding the correction quantity to the current output of the controller to generate a final control instruction; S9, constructing an extended state observer, wherein the extended state observer takes actual displacement and control instructions of the system as input, and estimates and outputs the total disturbance of the system in real time; The total disturbance estimation value output by the extended state observer is used for performing feedforward compensation on the control instruction so as to inhibit random high-frequency forging impact interference; the online iterative learning algorithm in the step S8 is used for compensating and correcting accumulated track deviation caused by axial plastic extension of the workpiece.
2. The method for correcting and controlling the track of the multi-axis electro-hydraulic servo system of the radial forging machine according to claim 1, wherein in the step S2, the nonlinear state space equation model of the valve-controlled hydraulic cylinder is expressed as: Wherein: the unit is m for the displacement of the piston rod; The unit is m/s for the movement speed of the piston rod; the unit is Pa for the pressure of the rod cavity of the hydraulic cylinder; the unit is Pa, which is the pressure of a rodless cavity of the hydraulic cylinder; The unit is m/s for the speed of the piston rod; The unit is m/s 2 for the acceleration of the piston rod; the unit is Pa/s for the pressure change rate of the rod cavity; the pressure change rate of the rodless cavity is Pa/s; the unit is kg; the unit of the input voltage is V; The area of the piston with the rod cavity is m 2 ; the area of the piston is the area of a rodless cavity, and the unit is m 2 ; The unit is N, which is the external load force; the volume of the rodless cavity is m 3 ; The volume of the rod cavity is m 3 ; Is the square root term of the equivalent pressure of the flow coefficient of the flow of the rod cavity, and the unit is ; Is the equivalent pressure square root term of the rodless cavity side flow coefficient, and is expressed in units of ; The unit is m 3 /(s.Pa) which is the internal leakage coefficient; Is the flow coefficient, the unit is ; The unit is m/V for servo valve gain; the hydraulic oil elastic modulus is Pa; The unit is m for the displacement of the valve core; the unit is Pa for the pressure of the hydraulic oil source.
3. The method for correcting and controlling the track of the multi-axis electro-hydraulic servo system of the radial forging machine according to claim 2, wherein the standard nonlinear state space equation model is expressed as: Wherein: Is an open loop system state variable; Is the derivative of the open loop system state variable; Is an observed value; Is an open loop system state matrix; is an input matrix; Is an output matrix.
4. The method for correcting and controlling the track of the multi-axis electro-hydraulic servo system of the radial forging machine according to claim 1, wherein in the step S3, the linearization process is specifically that the electro-hydraulic servo system is controlled to execute a constant motion test, rodless cavity pressure, rod cavity pressure and servo valve core displacement data of a hydraulic cylinder in the stage are collected and calculated to be arithmetic average value, the arithmetic average value is used as a steady-state working point parameter to calculate a constant input matrix B after linearization, and the rodless cavity volume which varies with the piston displacement in the system matrix A is calculated With rod chamber volume Using initial volume of rodless chamber with piston in intermediate position of stroke Initial volume of rod cavity Substitution is performed.
5. The method for correcting and controlling the trajectory of a multiaxial electro-hydraulic servo system of a radial forging machine according to claim 1, wherein in step S4, the closed-loop transfer function is as follows The method comprises the following steps: firstly, carrying out Laplacian transformation on a linearization state space equation to obtain an open-loop transfer function : Wherein: laplace transform for the output value; Laplacian transformation for an input value; is a system matrix; is an input matrix; Is an output matrix; Is a unit matrix; Is a Laplace transform operator; Selecting sampling step length Will open loop transfer function Conversion to discrete open loop transfer function : Wherein: Is that A transformation operator; Is a z transform operator; transfer function combined with PID controller : Finally, a closed loop transfer function is obtained : Wherein: is a proportionality coefficient; is an integral coefficient; Is a differential coefficient; Is a tracking error.
6. The method for trajectory modification and control of a multiaxial electro-hydraulic servo of a radial forging machine according to claim 1, wherein in step S5, the discrete state space equation is expressed as: Wherein: A discrete state variable of the closed loop system at the moment k; Observing a variable at the moment k for a closed loop system; the observation variable of the closed loop system at the moment k; Is a closed loop system state matrix; is an input matrix; Is an output matrix.
7. The method for correcting and controlling the track of the multi-axis electro-hydraulic servo system of the radial forging machine according to claim 1, wherein in step S7, a recurrence method is adopted for future generation Predicting the system state of each sampling step length, wherein a prediction formula is as follows: Wherein: Is that Predicting value of system state variable at moment; For the current moment Acquiring actual system state variables in real time; To at the same time Ideal tracking track at moment; Is that A system disturbance predicted value at a moment; , to predict the number of sampling steps in the time domain.
8. The method for trajectory correction and control of a multi-axis electro-hydraulic servo system of a radial forging machine according to claim 1, wherein in step S8, the process of calculating the trajectory correction by the online iterative learning algorithm specifically comprises: Future based on prediction Prediction output of individual sampling steps Obtaining a predicted tracking error: Wherein: To the future (future) Predictive tracking errors for the individual sampling steps; To at the same time Ideal tracking track at moment; For the current moment Acquiring actual system state variables in real time; Is that A system disturbance predicted value at a moment; Is the system disturbance at the kth moment; Is a closed loop system state matrix; is an input matrix; Is an output matrix; Is a matrix A kind of electronic device A power of the second; obtaining the first step by adopting an online iterative learning algorithm Secondary trajectory correction amount: Wherein: To be at future time Track correction is performed A secondary trajectory correction amount; gain coefficients for the trajectory correction; correction of the track Superimposed to the system at the present time Input trajectory of time of day And obtaining a corrected system input track: Wherein: Inputting a track for the corrected system; The trajectory is input for the system before correction.
9. The method for correcting and controlling the track of the multi-axis electro-hydraulic servo system of the radial forging machine according to claim 8, wherein the gain factor is The value of (2) is to balance the convergence speed and convergence stability and predict the time domain The value of (2) needs to cover the influence period of single impact interference.
10. The method for trajectory modification and control of a multiaxial electro-hydraulic servo system of a radial forging machine according to claim 1, wherein in step S9, the mathematical model of the extended state observer is: Wherein: is an observation error; is the actual displacement of the system; 、 And Is an observer state variable; 、 And Is the derivative of the observer state variable; 、、、 And The coefficient to be set is; is the actual input of the system; the function is a piecewise function: Wherein: is a function variable; Is a shape factor; Is a transition bandwidth coefficient; As a sign function.

Description

Track correction and control method for multi-axis electro-hydraulic servo system of radial forging machine Technical Field The invention belongs to the technical field of electrohydraulic servo control, and particularly relates to a track correction and control method of a radial forging machine multi-axis electrohydraulic servo system integrating expanded state observation and online iterative learning. Background Radial forging is used as a key metal plastic processing technology, and continuous local plastic deformation is generated on a metal blank through the synergistic effect of multidirectional radial concentrated force, so that high-precision, high-strength and high-performance shaft, pipe and special-shaped parts are finally obtained. The technology is widely applied to precision forming of high-end equipment core components such as main shafts of aeroengines, transmission shafts of heavy vehicles, nuclear power main pipelines and the like, and is one of key processes indispensable to modern manufacturing industry. The core actuating mechanism of the radial forging machine generally comprises four shaft systems, namely a hammer head shaft, a position shaft, a clamping pressure shaft, a drawing mandrel and the like, and each shaft system is driven by a multi-shaft electrohydraulic servo system to realize highly coordinated precise movement. The planning and control precision of the motion trail directly determines the geometric dimension, shape precision, internal organization and mechanical performance of the forging piece, and is a core link for guaranteeing the quality of the final product. In order to meet the requirement of high-precision forging, a multi-shaft closed-loop electrohydraulic servo control system is commonly adopted in a modern radial forging machine. The track planning and tracking control serve as core functions of the system, are responsible for generating ideal motion tracks of all shafting according to process requirements, and ensure that actual motion accurately reproduces the tracks through real-time feedback control. However, the radial forging process is extremely harsh and is mainly characterized in the following two aspects. (1) High frequency, high energy random forging impact, in which hundreds or even thousands of forging actions per minute produce a significant impact load in millisecond-scale time. The load is transmitted to the whole electrohydraulic servo system through the frame and the hydraulic pipeline, and causes strong pressure pulsation, vibration and nonlinear transient response of the valve control cylinder system, thereby forming dynamic interference with large amplitude, wide frequency bandwidth and strong randomness. (2) Cumulative perturbation induced by axial plastic elongation of the workpiece as forging proceeds, significant axial plastic flow of the metal blank occurs, resulting in a continuous increase in its length. The load change caused by the lengthening of the workpiece is a deterministic slow-changing process, but the additional load generated by the position shaft often exceeds the clamping force far, and if targeted track compensation is not carried out, tracking errors are accumulated continuously, so that the size consistency is seriously affected. In the face of the complex interference, the conventional control and trajectory planning method has significant limitations. In the track planning aspect, the conventional method is mostly static preset, such as an S-shaped curve, a polynomial curve and the like, so as to ensure smooth track, but real-time dynamic interference in the forging process is not considered. The essence of the method is an open loop feedforward plan, which cannot be adaptively adjusted according to the actual running state (such as the impact and the load change) of the system, so that the preset ideal track is difficult to accurately track under the actual high-frequency impact working condition, and the tracking error is obvious along with the fluctuation of the working condition. At the control strategy level, a single advanced control algorithm is difficult to effectively cope with the interference with the two different characteristics at the same time. (1) The disturbance rejection control for random impact is that a model-free self-adaptive control method based on an RBF neural network disturbance observer disclosed in Chinese patent application with publication number of CN112925208A and an intelligent self-learning PID control method disclosed in Chinese patent application with publication number of CN113110037A are capable of estimating and compensating unknown disturbance to a certain extent by using the data-driven or model-independent method, but the design is initially biased towards the treatment of parameter uncertainty and general external load disturbance. In the random impact context of radial forging up to hundreds of hertz, the dynamic response speed and observation bandwidth of such method