An adaptive pressure control method and control system for a machine tool hydraulic clamping system

Abstract

This invention provides an adaptive pressure control method and control system for a machine tool hydraulic clamping system, relating to the field of machine tool hydraulic control technology. The method includes: real-time acquisition of the pressure in both chambers of the clamping cylinder, piston rod displacement, and oil temperature; dynamic determination of the target clamping pressure based on the workpiece material and machining conditions; construction of an RBF neural network for online estimation of unmodeled dynamics; solving an improved robust control law using the estimated values ​​to obtain the control quantity; and outputting the control signal to an electro-hydraulic servo valve to achieve real-time adaptive control. This invention, through a composite algorithm of RBF neural network and improved robust control, effectively compensates for the nonlinearity and parameter uncertainties of the hydraulic system, achieving stable and reliable control of the clamping pressure and preventing workpiece deformation or loosening.

Description

Technical Field

[0001] This invention relates to the field of machine tool hydraulic control technology, and in particular to an adaptive pressure control method and control system for a machine tool hydraulic clamping system. Background Technology

[0002] Hydraulic clamping systems are crucial devices for securing workpieces during CNC machine tool machining. They utilize hydraulic power to drive an actuator, generating clamping force to ensure the workpiece remains stable during cutting, preventing deviations in machining accuracy or safety accidents due to displacement. These systems are widely used in various metal cutting machine tools, particularly machining centers and milling machines, where reliable workpiece fixation is essential.

[0003] Currently, most machine tool hydraulic clamping systems on the market still rely on traditional fixed-parameter control algorithms for pressure control, with proportional-integral-derivative (PID) control and its improved forms being the most common. The basic logic of this type of control scheme is as follows: the operator presets a fixed clamping pressure value based on experience; the system collects the pressure signal from a specific chamber of the clamping cylinder through a pressure sensor; the deviation between the collected actual pressure and the preset pressure is calculated; and then feedback adjustment is performed based on fixed proportional, integral, and derivative parameters to achieve closed-loop control of the clamping force. In practical applications, once the control parameters of this type of system are set, they remain unchanged throughout the entire machining process and cannot be automatically adjusted according to changes in the machining scenario.

[0004] However, as modern machining develops towards higher precision, higher efficiency, more variety, and smaller batch production, the limitations of the aforementioned traditional control technologies are becoming increasingly apparent, specifically in the following aspects:

[0005] First, the control algorithm has poor adaptability and struggles to match complex and ever-changing machining conditions. Traditional control algorithms are designed based on fixed parameters and do not fully consider changes in working conditions such as workpiece material switching, cutting force fluctuations, and cutting speed variations that may occur during machining. When the workpiece changes from one material to another, the required clamping force often differs significantly, but traditional systems cannot dynamically adjust the clamping pressure parameters, leading to a decrease in control accuracy during multiple working condition switching and making it difficult to meet diverse machining needs.

[0006] Second, the nonlinearity and disturbance compensation capabilities are severely insufficient. Hydraulic systems inherently possess nonlinear characteristics, such as the viscosity-temperature properties of hydraulic oil, nonlinear friction, and nonlinear valve flow. Simultaneously, the system also experiences parameter fluctuations, such as unmodeled dynamics like the change in the hydraulic oil's bulk modulus with temperature and the change in the leakage coefficient with pressure. Traditional control algorithms lack effective compensation mechanisms for these nonlinearities and uncertainties. When these factors change, the system is prone to pressure fluctuations, which in extreme cases may lead to excessive clamping force causing workpiece deformation, or insufficient clamping force causing workpiece loosening, seriously affecting processing quality and production safety.

[0007] Third, the sensing and detection dimensions are limited, resulting in incomplete system status perception. Existing systems mostly rely on a single pressure sensor to collect signals, lacking real-time monitoring of key operating parameters such as piston rod displacement and hydraulic oil temperature. Displacement information can be used to determine whether clamping is in place or whether the clamping cylinder has reached its stroke limit; oil temperature information directly affects key parameters such as hydraulic oil viscosity and elastic modulus. Due to the lack of perception of this multi-dimensional information, the system cannot comprehensively reflect its operating status, making it difficult to provide sufficient data support for control decisions and to promptly detect potential faults.

[0008] Fourth, the safety protection mechanism is inadequate, and the system reliability is insufficient. Existing systems largely rely on relief valves for passive overload protection, which is slow to respond and is reactive. Traditional systems lack proactive early warning and rapid response mechanisms for abnormal conditions such as insufficient pressure or hydraulic oil overheating. In the event of sudden malfunctions, such as sensor failure or valve core jamming, the system often fails to respond promptly, easily leading to workpiece damage, equipment failure, or even safety accidents. The system's operational reliability needs improvement.

[0009] To address the aforementioned issues, it is necessary to develop a novel adaptive pressure control technology for machine tool hydraulic clamping systems to improve the system's adaptability to operating conditions, anti-interference capabilities, state awareness, and safety and reliability. Summary of the Invention

[0010] To address the shortcomings of existing technologies, this invention provides an adaptive pressure control method and control system for machine tool hydraulic clamping systems. It has advantages such as multi-source sensor fusion, adaptive pressure regulation, nonlinearity and interference compensation, and active safety protection, and solves the problems of poor working condition adaptability, pressure fluctuation, poor clamping effect, and insufficient safety and reliability in the current pressure control technology of machine tool hydraulic clamping systems.

[0011] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0012] An adaptive pressure control method for a machine tool hydraulic clamping system includes the following steps:

[0013] Data Acquisition: Real-time acquisition of the rodless chamber pressure P1, rod chamber pressure P2, and piston rod displacement x of the double-acting clamping cylinder in the machine tool hydraulic clamping system. p and hydraulic oil temperature T;

[0014] Based on preset workpiece material parameters and machining condition parameters, the target clamping pressure P under the current working condition is dynamically determined by querying a preset pressure mapping table. target ;

[0015] An RBF neural network is constructed, using the collected data as input, to perform online estimation of the unmodeled dynamics of the hydraulic system, obtaining the estimated values ​​of the unmodeled dynamics; the unmodeled dynamics include at least the hydraulic oil leakage Δq and the nonlinear friction force F. f ;

[0016] According to the target clamping pressure P target The pressure tracking error e is constructed by comparing the actual clamping pressure calculated from the rodless chamber pressure P1 and the rod chamber pressure P2. Combined with the obtained unmodeled dynamic estimate, the improved robust control law is solved to obtain the control law U. The improved robust control law includes an adaptive model compensation term U for compensating for known and unmodeled dynamics of the system. a And a robust feedback term U for suppressing parameter uncertainty. s ;

[0017] According to the control law U, a control signal is output to the electro-hydraulic servo valve to adjust the valve core displacement, thereby controlling the actual clamping pressure of the double-acting clamping cylinder to reach the target clamping pressure P. target .

[0018] Furthermore, the collected rodless chamber pressure P1, rod chamber pressure P2, and piston rod displacement x were analyzed. p The hydraulic oil temperature T and the pressure signal are filtered using a first-order low-pass filter to obtain the filtered rodless chamber pressure signal P1', the filtered rod chamber pressure signal P2', and the filtered piston rod displacement x. p The filtered hydraulic oil temperature T is used as input data for the RBF neural network.

[0019] Furthermore, the RBF neural network has a three-layer structure, including an input layer, a hidden layer, and an output layer; the basis functions of the hidden layer are Gaussian kernel functions, and their expression is:

[0020] ,

[0021] Where: x is the input vector, c j Let b be the center vector of the j-th kernel function. j Let be the width parameter of the j-th kernel function;

[0022] The RBF neural network is trained online using the gradient descent method, and the adjustment formulas for its weights, width parameter, and center vector are as follows:

[0023] Weight adjustment formula: ,

[0024] ;

[0025] Width parameter adjustment formula: ;

[0026] Central vector adjustment formula: ,

[0027] ;

[0028] In the formula:

[0029] The number of iterations; Neurons in the hidden layer; The dimension of the input vector; This is the system output for step n; This is the network output for step n; This is the output of the hidden layer at step n; The learning rate is the weight. The learning rate is the width parameter; Learning rate for the center vector; The weight momentum factor; The momentum factor is the width parameter. The momentum factor of the center vector; This is the weight adjustment amount for the nth step; The weight is for the (n+1)th step; This is the adjustment amount for the width parameter in step n; This is the width parameter for the nth step; For the nth step Input feature values ​​of each hidden layer neuron; For the nth step The center vector of each hidden layer neuron; For the nth step The center vector of the nth hidden layer neuron is... Adjustment amount for each dimension vector; For the first Step 1 The first hidden layer neuron center vector of the _ ... Each dimension is a vector value.

[0030] Furthermore, the improved robust control law U is expressed as: ;

[0031] Among them, U a The adaptive model compensation term is represented as:

[0032] ,

[0033] in: , , Elements of an uncertain parameter matrix; The bulk modulus of hydraulic oil; This is the internal leakage coefficient; The external leakage coefficient; The output of the RBF neural network is the unmodeled dynamic estimate. For the target displacement;

[0034] f2, f3, and f4 are functions related to the system state, as detailed below:

[0035] ;

[0036] ;

[0037] ;

[0038] In the formula: , The volume of the two chambers of the hydraulic cylinder; Where A is the load pressure; A is the effective working area of ​​the piston; This represents the displacement of the piston rod.

[0039] U s For robust feedback items, it is represented as: ,

[0040] Where k1 is the linear feedback gain; e is the pressure tracking error; h1 is the upper bound of the unmodeled dynamic estimation error; and δ is the preset positive design parameter.

[0041] Furthermore, the formula for calculating the pressure tracking error e is as follows: .

[0042] Furthermore, the pressure mapping table is calibrated using the following empirical formula for milling force:

[0043] ,

[0044] In the formula, F c Main cutting force; C F a is the milling force coefficient; p For the depth of cut; a f For feed rate; a wd is the cutting width; d0 is the tool diameter; n is the spindle speed; xF, yF, uF, qF, wF are exponential coefficients; k Fc The milling force correction coefficient is used; both the exponential coefficient and the milling force correction coefficient are obtained by consulting the machining process manual.

[0045] A control system based on the adaptive pressure control method of the machine tool hydraulic clamping system, comprising:

[0046] A hydraulic power module is used to provide hydraulic oil with adjustable pressure and flow.

[0047] The clamping execution module includes a double-acting clamping cylinder, wherein the rodless chamber and the rod chamber of the double-acting clamping cylinder are respectively connected to a hydraulic power module for performing clamping actions under the drive of hydraulic oil;

[0048] The sensing and detection module includes a pressure sensor, a displacement sensor, and a temperature sensor, which are used to collect the rodless chamber pressure P1, the rod chamber pressure P2, the piston rod displacement x, and the hydraulic oil temperature T of the double-acting clamping cylinder in real time, respectively.

[0049] The control processing module is electrically connected to the signal output terminal of the sensing and detection module and the control input terminal of the hydraulic power module, respectively. The control processing module is used to execute the adaptive pressure control method.

[0050] The safety protection module is bidirectionally electrically connected to the control processing module and is used to perform protective actions under abnormal operating conditions.

[0051] Furthermore, the control processing module includes an embedded controller based on a field-programmable gate array (FPGA) and control software; the FPGA chip is integrated inside the embedded controller, and the control software is deployed in the embedded controller to implement the composite adaptive pressure control algorithm; the FPGA chip has a clock frequency of not less than 40MHz and is internally configured with a first-in-first-out (FIFO) data buffer queue with a depth of not less than 1024 bytes and a bit width of not less than 32 bits to process multi-channel sensor data collected by the sensing and detection module.

[0052] Furthermore, the control software includes:

[0053] The data acquisition module is used to acquire the signal output by the sensing and detection module through the analog input interface of the embedded controller;

[0054] An algorithm computation module, deployed inside the field-programmable gate array chip, is used to execute the computation of the composite adaptive pressure control algorithm in parallel;

[0055] The human-computer interaction module provides an interface for parameter configuration and status monitoring.

[0056] The data storage module is used to periodically store system operation data.

[0057] Furthermore, the security protection module includes:

[0058] A pressure relay is connected to the hydraulic lines of the rodless chamber and the rod chamber of the double-acting clamping cylinder to monitor the pressure of the two chambers in real time.

[0059] An electromagnetic shut-off valve is installed in series on the hydraulic pipeline between the hydraulic power module and the double-acting clamping cylinder;

[0060] The signal output terminal of the pressure relay is electrically connected to the signal input terminal of the control processing module, and the control output terminal of the control processing module is electrically connected to the electronic control terminal of the electromagnetic shut-off valve.

[0061] The beneficial effects of this invention are as follows:

[0062] 1. The adaptive pressure control method and control system of the machine tool hydraulic clamping system of the present invention achieves adaptive decision-making of the target pressure by pre-setting a pressure mapping table based on measured data and dynamically determining the target clamping pressure according to workpiece material parameters and machining condition parameters. During machining, the most suitable target clamping pressure can be automatically matched according to information such as workpiece hardness, elastic modulus, and current cutting force and cutting speed, without manual intervention. At the same time, by combining an RBF neural network with an improved robust control composite algorithm, the control parameters are adjusted in real time, enabling the system to effectively adapt to complex working condition changes such as workpiece material switching and cutting force fluctuations, significantly improving the control accuracy under multiple working conditions.

[0063] 2. The adaptive pressure control method and control system for the machine tool hydraulic clamping system described in this invention, by constructing an RBF neural network and using multi-source sensor data as input, accurately estimates the unmodeled dynamics of the hydraulic system (including hydraulic oil leakage, nonlinear friction, etc.) online, solving the problem of difficulty in modeling system nonlinearity. Simultaneously, an improved robust control law is designed, integrating the unmodeled dynamic estimates through an adaptive model compensation term and suppressing interference caused by parameter uncertainties (including changes in the hydraulic oil bulk modulus and leakage coefficient fluctuations) through a robust feedback term. The synergistic effect of these two mechanisms forms an effective nonlinearity and interference compensation mechanism, significantly suppressing the influence of hydraulic oil leakage, nonlinear friction, and changes in the elastic modulus on clamping pressure, achieving stable and reliable clamping pressure control, and avoiding workpiece deformation or loosening due to pressure fluctuations.

[0064] 3. The adaptive pressure control method and control system for the machine tool hydraulic clamping system described in this invention solves the problem of incomplete state perception in traditional systems by setting up a sensing and detection module including a pressure sensor, a displacement sensor, and a temperature sensor to collect multi-dimensional data in real time, such as the pressure in the rodless and rod chambers of the double-acting clamping cylinder, the piston rod displacement, and the hydraulic oil temperature. The collected data is filtered by a first-order low-pass filter before being transmitted to the control processing module, providing comprehensive, accurate, and reliable real-time operating data for control decisions.

[0065] 4. The adaptive pressure control method and control system of the machine tool hydraulic clamping system described in this invention, by setting up a safety protection module including a pressure relay and an electromagnetic shut-off valve, and employing a hardware interlock design, achieves proactive early warning and rapid handling of abnormal working conditions. When the pressure exceeds preset upper and lower limits or the hydraulic oil temperature exceeds a threshold, the system can quickly trigger an emergency stop, close the electromagnetic shut-off valve, cut off the hydraulic pipeline, control the electro-hydraulic servo valve to return to the neutral position, issue an alarm signal, or start the cooling system through a hardware interrupt signal, achieving multi-level and multi-dimensional proactive safety protection. The hardware interlock design ensures that the emergency stop signal has higher priority than the conventional control signal, and can reliably execute protective actions even in the event of controller software malfunction, fundamentally solving the problems of passive protection and slow response in traditional systems, and significantly improving the reliability and safety of system operation. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings described below are some embodiments of the present invention. For those skilled in the art, it is obvious that other drawings can be obtained from these drawings without creative effort.

[0067] Figure 1 This is a schematic diagram of the adaptive pressure control system of the machine tool hydraulic clamping system described in this invention.

[0068] Figure 2 This is a flowchart of the adaptive pressure control method for the machine tool hydraulic clamping system described in this invention.

[0069] Figure 3 This is a schematic diagram of the security protection logic of the present invention.

[0070] Figure 4 This is a schematic diagram of the RBF neural network structure of the present invention. Detailed Implementation

[0071] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0072] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "axial," "radial," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0073] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0074] like Figure 2 As shown, the adaptive pressure control method for the machine tool hydraulic clamping system of the present invention includes the following steps:

[0075] S1: System initialization. Set the structure of the RBF neural network, determine the number of nodes in its input layer, hidden layer, and output layer, and select a Gaussian function as the basis function for the hidden layer; at the same time, set the coefficients of the linear feedback term in the improved robust control law.

[0076] S2: Data Acquisition. Real-time acquisition of the rodless chamber pressure P1, rod chamber pressure P2, and piston rod displacement x of the double-acting clamping cylinder in the machine tool hydraulic clamping system. p and hydraulic oil temperature T;

[0077] S3: Data Processing. A first-order low-pass filter is used to filter the acquired signals to obtain the filtered rodless cavity pressure signal P1', the filtered rod cavity pressure signal P2', and the filtered piston rod displacement x. p The filtered hydraulic oil temperature T is used as input data for the RBF neural network.

[0078] S4: Based on the preset workpiece material parameters and machining condition parameters, dynamically determine the target clamping pressure P under the current working condition by querying the preset pressure mapping table. target Workpiece material parameters include hardness. and elastic modulus Machining parameters include cutting force and cutting speed This pressure mapping table is pre-established and stored based on a large amount of clamping force experimental data from various workpieces of different materials under different milling process parameters.

[0079] The pressure mapping table is calibrated using the following empirical formula for milling force:

[0080] ,

[0081] In the formula, F c Main cutting force; C F a is the milling force coefficient; p For the depth of cut; a f For feed rate; a w d is the cutting width; d0 is the tool diameter; n is the spindle speed; xF, yF, uF, qF, wF are exponential coefficients; k Fc The milling force correction coefficient is used; both the exponential coefficient and the milling force correction coefficient are obtained by consulting the machining process manual.

[0082] S5: To compensate for the inherent nonlinearity and unknown disturbances in the hydraulic system, an RBF neural network is constructed to estimate the unmodeled dynamics of the hydraulic system, including hydraulic oil leakage. and nonlinear friction Filtered pressure , Displacement ,temperature .like Figure 4 As shown, the RBF neural network has a three-layer structure, including an input layer, a hidden layer, and an output layer; the input vector of this network is the filtered system state. Its output is an unmodeled dynamic estimate. ,in The estimated values ​​of the weight matrix for the RBF neural network. Let be the Gaussian kernel function vector. The basis functions of the hidden layer adopt the Gaussian kernel function, and its expression is:

[0083] ,

[0084] Where: x is the input vector, c j Let b be the center vector of the j-th kernel function. j Let be the width parameter of the j-th kernel function;

[0085] S6: Online Training of the Neural Network. The RBF neural network is trained online using the gradient descent method. The adjustment formulas for its weights, width parameter, and center vector are as follows:

[0086] Weight adjustment formula: ,

[0087] ;

[0088] Width parameter adjustment formula: ;

[0089] Central vector adjustment formula: ,

[0090] ;

[0091] In the formula:

[0092] The number of iterations; Neurons in the hidden layer; The dimension of the input vector; This is the system output for step n; This is the network output for step n; This is the output of the hidden layer at step n; The learning rate is the weight. The learning rate is the width parameter; Learning rate for the center vector; The weight momentum factor; The momentum factor is the width parameter. The momentum factor of the center vector; This is the weight adjustment amount for the nth step; The weight is for the (n+1)th step; This is the adjustment amount for the width parameter in step n; This is the width parameter for the nth step; For the nth step Input feature values ​​of each hidden layer neuron; For the nth step The center vector of each hidden layer neuron; For the nth step The center vector of the nth hidden layer neuron is... Adjustment amount for each dimension vector; For the first Step 1 The first hidden layer neuron center vector of the _ ... Each dimension is a vector value.

[0093] S7: According to the target clamping pressure P target The pressure tracking error e is constructed by comparing the actual clamping pressure calculated from the rodless chamber pressure P1 and the rod chamber pressure P2. Combined with the obtained unmodeled dynamic estimate, the improved robust control law is solved to obtain the control law U. The improved robust control law includes an adaptive model compensation term U for compensating for known and unmodeled dynamics of the system. a And a robust feedback term U for suppressing parameter uncertainty. s ;

[0094] The improved robust control law U is expressed as: ;

[0095] Among them, U a The adaptive model compensation term is represented as:

[0096] ,

[0097] in: , , Elements of an uncertain parameter matrix; The bulk modulus of hydraulic oil; This is the internal leakage coefficient; The external leakage coefficient; The output of the RBF neural network is the unmodeled dynamic estimate. For the target displacement;

[0098] f2, f3, and f4 are functions related to the system state, as detailed below:

[0099] ;

[0100] ;

[0101] ;

[0102] In the formula: , The volume of the two chambers of the hydraulic cylinder; Where A is the load pressure; A is the effective working area of ​​the piston; This represents the displacement of the piston rod.

[0103] Us For robust feedback items, it is represented as: ,

[0104] Where k1 is the linear feedback gain; e is the pressure tracking error; h1 is the upper bound of the unmodeled dynamic estimation error; and δ is the preset positive design parameter.

[0105] The formula for calculating the pressure tracking error e is: .

[0106] S8: Based on the obtained control law U, output the corresponding analog or digital control signal to the electro-hydraulic servo valve to adjust the valve core displacement, thereby controlling the actual clamping pressure of the double-acting clamping cylinder to reach the target clamping pressure P. target Repeat steps S2 to S8 to begin a new cycle of data acquisition, estimation, and control, achieving real-time, adaptive pressure closed-loop control.

[0107] To verify the stability of the algorithm, a positive definite Lyapunov function is constructed. ,in The adaptive gain matrix is ​​obtained by differentiating it with respect to V. When t→∞, e→0, proving that the system has asymptotic stability, ensuring that the algorithm maintains stable control performance throughout long-term operation.

[0108] like Figure 1 As shown, the adaptive pressure control system of the machine tool hydraulic clamping system of the present invention includes a hydraulic power module, a clamping execution module, a sensing and detection module, a control processing module, and a safety protection module.

[0109] The hydraulic power module is connected to the clamping execution module via hydraulic lines, providing the clamping execution module with hydraulic oil of adjustable pressure and flow. Specifically, the hydraulic power module includes an oil tank, a hydraulic pump, a check valve, a relief valve, and an electro-hydraulic servo valve, connected sequentially via oil lines. The oil inlet of the hydraulic pump is connected to the oil tank, and the oil outlet is connected to the oil inlet of the electro-hydraulic servo valve via the check valve. The relief valve is connected in parallel between the oil outlet of the hydraulic pump and the oil tank to limit the maximum system pressure. The working ports A and B of the electro-hydraulic servo valve are connected to the clamping execution module via hydraulic lines.

[0110] The clamping execution module includes a double-acting clamping cylinder. The rodless chamber port and the rod chamber port of this double-acting clamping cylinder are respectively connected to the working ports A and B of the electro-hydraulic servo valve via hydraulic lines. When the piston rod of the clamping cylinder extends, it clamps the workpiece.

[0111] The sensing and detection module is used to collect the system's operating status parameters in real time. This module includes at least one pressure sensor, one displacement sensor, and one temperature sensor. Two pressure sensors are installed on the hydraulic lines connecting the rodless chamber and the rod chamber, respectively, to collect the pressure P1 in the rodless chamber and the pressure P2 in the rod chamber. The displacement sensor is installed at the piston rod end of the double-acting clamping cylinder to collect the displacement of the piston rod, which can be used to determine the clamping state. The temperature sensor is installed inside the oil tank to collect the real-time temperature T of the hydraulic oil. The signal output terminals of all sensors are electrically connected to the signal input terminals of the control and processing module.

[0112] The control processing module is the control core of this device, comprising an embedded controller based on a field-programmable gate array (FPGA) and control software built into the controller. The FPGA chip serves as the core computing unit and is integrated within the embedded controller. The control processing module incorporates a composite adaptive pressure control algorithm based on radial basis function neural networks and improved robust control, used to calculate and output control commands in real time based on data collected by the sensing module.

[0113] The safety protection module is used to quickly intervene in abnormal operating conditions to ensure the safety of equipment and workpieces. It includes pressure relays and solenoid shut-off valves. Two pressure relays are connected to the hydraulic lines of the rodless and rod-side chambers of the double-acting clamping cylinder, respectively, for real-time monitoring of the pressure in both chambers. The solenoid shut-off valve is normally open and is installed in series on the hydraulic line between the electro-hydraulic servo valve and the double-acting clamping cylinder (for example, it can be connected in series between port A of the electro-hydraulic servo valve and the rodless chamber). The signal output terminal of the pressure relay is electrically connected to the signal input terminal of the control processing module, and the control output terminal of the control processing module is electrically connected to the electrical control terminal of the solenoid shut-off valve. Furthermore, the safety protection module and the control processing module employ a hardware-interlocked bidirectional electrical connection design to ensure the highest priority of the emergency stop signal.

[0114] During operation, the control processing module dynamically calculates the required target clamping pressure based on the preset adaptive pressure control algorithm and the real-time status data fed back by the sensor detection module. It then outputs a corresponding current signal to drive the electro-hydraulic servo valve. By controlling the opening and direction of the valve core, it precisely adjusts the oil entering the rodless and rod chambers of the double-acting clamping cylinder, thereby achieving adaptive and high-precision control of the clamping pressure.

[0115] The core of the control processing module is an FPGA chip, which adopts a programmable logic array architecture. To ensure the real-time performance of the algorithm, its internal global clock frequency is set to no less than 40MHz. For high-speed transmission of multi-channel sensor data, a data buffer queue based on the FIFO (First-In-First-Out) principle is configured inside the FPGA. This queue has a depth of no less than 1024 bytes and a data bit width of no less than 32 bits to match the data throughput rate between high-speed AD sampling and parallel processing units.

[0116] The control software adopts a modular design and includes at least a data acquisition module, an algorithm calculation module, a human-computer interaction module, and a data storage module.

[0117] Data Acquisition Module: This module connects to each sensor via the analog input interface at the bottom layer of the embedded controller. The hardware sampling rate of the analog input interface is no less than 500kS / s, and the resolution is no less than 12 bits to ensure the accuracy and timeliness of the acquired signals.

[0118] Algorithm Calculation Module: This module is the core carrier of the algorithm and is configured to be implemented as hardware logic circuits inside the FPGA chip. This means that the complex calculations such as RBF neural network training and improved robust control law solving described in Example 2 will be executed in a parallel pipeline manner, greatly shortening the control cycle.

[0119] Human-Machine Interaction Module: This module runs on an embedded processor connected to the FPGA and is responsible for providing a parameter configuration interface and a status monitoring interface. Operators can use this module to input parameters such as workpiece material and processing conditions, and view key data such as pressure and displacement in real time.

[0120] Data storage module: This module is connected to the algorithm calculation module and is used to periodically record system operation data. For ease of later analysis and processing, the data is stored in CSV (comma-separated values) file format on local storage media, and the storage period can be set to once per minute.

[0121] like Figure 3 As shown, the safety protection module, through the hardware interlock design of pressure relays, electromagnetic shut-off valves, and control processing modules, achieves accurate identification and protective control of critical abnormal conditions such as abnormal pressure and hydraulic oil overheating, comprehensively improving the reliability and safety of system operation. The specific working logic is as follows:

[0122] The control processing module presets an upper pressure limit. Lower limit of pressure The hydraulic oil temperature threshold of 80℃ is used as a safety protection criterion, and the pressure of the rodless chamber of the double-acting clamping cylinder transmitted in real time by the sensor detection module is also monitored. The hydraulic oil temperature T is measured and compared in real time.

[0123] When the pressure in the rodless chamber exceeds the preset safety threshold range, i.e. when or Upon receiving the trigger signal, the pressure relay immediately sends a signal to the control processing module via electrical connection. The control processing module then immediately outputs a hardware emergency stop signal, which is transmitted to the solenoid shut-off valve via electrical connection. This drives the solenoid shut-off valve to quickly close the hydraulic line, cutting off the supply of hydraulic oil to the double-acting clamping cylinder. Simultaneously, it controls the electro-hydraulic servo valve to switch to the neutral position, stopping valve core displacement adjustment and preventing further abnormal pressure fluctuations. Furthermore, the control processing module simultaneously issues an alarm signal to prompt personnel to troubleshoot the problem.

[0124] When the hydraulic oil temperature exceeds the preset threshold (T>80℃), the control processing module sends a stop signal to the hydraulic pump in the hydraulic power module via electrical connection. At the same time, the cooling system is activated to cool the hydraulic oil. During the operation of the cooling system, the temperature sensor continuously collects the hydraulic oil temperature and transmits it to the control processing module. When the hydraulic oil temperature drops below the safe recovery threshold (T<60℃), the control processing module sends a signal to the hydraulic pump again to restart the hydraulic pump.

[0125] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0126] The detailed descriptions listed above are merely specific illustrations of feasible embodiments of the present invention and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. An adaptive pressure control method for a machine tool hydraulic clamping system, characterized in that, Includes the following steps: Data Acquisition: Real-time acquisition of the rodless chamber pressure P1, rod chamber pressure P2, and piston rod displacement x of the double-acting clamping cylinder in the machine tool hydraulic clamping system. p and hydraulic oil temperature T; Based on preset workpiece material parameters and machining condition parameters, the target clamping pressure P under the current working condition is dynamically determined by querying a preset pressure mapping table. target ; An RBF neural network is constructed, using the collected data as input, to perform online estimation of the unmodeled dynamics of the hydraulic system, obtaining the estimated values ​​of the unmodeled dynamics; the unmodeled dynamics include at least the hydraulic oil leakage Δq and the nonlinear friction force F. f ; According to the target clamping pressure P target The pressure tracking error e is constructed by comparing the actual clamping pressure calculated from the rodless chamber pressure P1 and the rod chamber pressure P2. By combining the obtained unmodeled dynamic estimates, the improved robust control law is solved to obtain the control law U; the improved robust control law includes an adaptive model compensation term U for compensating for the known dynamics and unmodeled dynamics of the system. a And a robust feedback term U for suppressing parameter uncertainty. s ; According to the control law U, a control signal is output to the electro-hydraulic servo valve to adjust the valve core displacement, thereby controlling the actual clamping pressure of the double-acting clamping cylinder to reach the target clamping pressure P. target .

2. The adaptive pressure control method for the machine tool hydraulic clamping system according to claim 1, characterized in that, The collected data include rodless chamber pressure P1, rod chamber pressure P2, and piston rod displacement x. p The hydraulic oil temperature T and the pressure signal are filtered using a first-order low-pass filter to obtain the filtered rodless chamber pressure signal P1', the filtered rod chamber pressure signal P2', and the filtered piston rod displacement x. p The filtered hydraulic oil temperature T is used as input data for the RBF neural network.

3. The adaptive pressure control method for the machine tool hydraulic clamping system according to claim 1, characterized in that, The RBF neural network has a three-layer structure, including an input layer, a hidden layer, and an output layer; the basis functions of the hidden layer are Gaussian kernel functions, and their expression is: ， Where: x is the input vector, c j Let b be the center vector of the j-th kernel function. j Let be the width parameter of the j-th kernel function; The RBF neural network is trained online using the gradient descent method, and the adjustment formulas for its weights, width parameter, and center vector are as follows: Weight adjustment formula: , ； Width parameter adjustment formula: ; Central vector adjustment formula: , ； In the formula: The number of iterations; Neurons in the hidden layer; The dimension of the input vector; This is the system output for step n; This is the network output for step n; This is the output of the hidden layer at step n; The learning rate is the weight. The learning rate is the width parameter; Learning rate for the center vector; The weight momentum factor; The momentum factor is the width parameter. The momentum factor of the center vector; This is the weight adjustment amount for the nth step; The weight is for the (n+1)th step; This is the adjustment amount for the width parameter in step n; This is the width parameter for the nth step; For the nth step Input feature values ​​of each hidden layer neuron; For the nth step The center vector of each hidden layer neuron; For the nth step The center vector of the nth hidden layer neuron is... Adjustment amount for each dimension vector; For the first Step 1 The first hidden layer neuron center vector of the _ ... Each dimension is a vector value.

4. The adaptive pressure control method for the machine tool hydraulic clamping system according to claim 1, characterized in that, The improved robust control law U is expressed as: ; Among them, U a The adaptive model compensation term is represented as: ， in: , , Elements of an uncertain parameter matrix; The bulk modulus of hydraulic oil; This is the internal leakage coefficient; The external leakage coefficient; The output of the RBF neural network is the unmodeled dynamic estimate. For the target displacement; f2, f3, and f4 are functions related to the system state, as detailed below: ； ； ； In the formula: , The volume of the two chambers of the hydraulic cylinder; Where A is the load pressure; A is the effective working area of ​​the piston; This represents the displacement of the piston rod. U s For robust feedback items, it is represented as: , Where k1 is the linear feedback gain; e is the pressure tracking error; h1 is the upper bound of the unmodeled dynamic estimation error; and δ is the preset positive design parameter.

5. The adaptive pressure control method for the machine tool hydraulic clamping system according to claim 4, characterized in that, The formula for calculating the pressure tracking error e is: .

6. The adaptive pressure control method for the machine tool hydraulic clamping system according to claim 1, characterized in that, The pressure mapping table is calibrated using the following empirical formula for milling force: ， In the formula, F c Main cutting force; C F a is the milling force coefficient; p For the depth of cut; a f For feed rate; a w d is the cutting width; d0 is the tool diameter; n is the spindle speed; xF, yF, uF, qF, wF are exponential coefficients; k Fc The milling force correction coefficient is used; both the exponential coefficient and the milling force correction coefficient are obtained by consulting the machining process manual.

7. A control system based on the adaptive pressure control method of the machine tool hydraulic clamping system according to any one of claims 1-6, characterized in that, include: A hydraulic power module is used to provide hydraulic oil with adjustable pressure and flow. The clamping execution module includes a double-acting clamping cylinder, wherein the rodless chamber and the rod chamber of the double-acting clamping cylinder are respectively connected to a hydraulic power module for performing clamping actions under the drive of hydraulic oil; The sensing and detection module includes a pressure sensor, a displacement sensor, and a temperature sensor, which are used to collect the rodless chamber pressure P1, the rod chamber pressure P2, the piston rod displacement x, and the hydraulic oil temperature T of the double-acting clamping cylinder in real time, respectively. The control processing module is electrically connected to the signal output terminal of the sensing and detection module and the control input terminal of the hydraulic power module, respectively. The control processing module is used to execute the adaptive pressure control method. The safety protection module is bidirectionally electrically connected to the control processing module and is used to perform protective actions under abnormal operating conditions.

8. The control system according to claim 7, characterized in that, The control processing module includes an embedded controller based on a field-programmable gate array (FPGA) and control software. The FPGA chip is integrated inside the embedded controller, and the control software is deployed in the embedded controller to implement the composite adaptive pressure control algorithm. The FPGA chip has a clock frequency of not less than 40MHz and is internally configured with a first-in-first-out (FIFO) data buffer queue with a depth of not less than 1024 bytes and a bit width of not less than 32 bits to process multi-channel sensor data collected by the sensing and detection module.

9. The control system according to claim 7, characterized in that, The control software includes: The data acquisition module is used to acquire the signal output by the sensing and detection module through the analog input interface of the embedded controller; An algorithm computation module, deployed inside the field-programmable gate array chip, is used to execute the computation of the composite adaptive pressure control algorithm in parallel; The human-computer interaction module provides an interface for parameter configuration and status monitoring. The data storage module is used to periodically store system operation data.

10. The control system according to claim 7, characterized in that, The security protection module includes: A pressure relay is connected to the hydraulic lines of the rodless chamber and the rod chamber of the double-acting clamping cylinder to monitor the pressure of the two chambers in real time. An electromagnetic shut-off valve is installed in series on the hydraulic pipeline between the hydraulic power module and the double-acting clamping cylinder; The signal output terminal of the pressure relay is electrically connected to the signal input terminal of the control processing module, and the control output terminal of the control processing module is electrically connected to the electronic control terminal of the electromagnetic shut-off valve.

Images