Low-carbon steel cold extrusion forming production line and forming dynamic pressure compensation process thereof

Abstract

The invention relates to the technical field of cold extrusion forming, in particular to a low-carbon steel cold extrusion forming production line and a forming dynamic pressure compensation process thereof. Comprising a feeding conveying line, a hydraulic machine body composed of an upper cross beam, a movable beam, a lower cross beam, a hydraulic system and a main cylinder, an extrusion mechanism, a finished product conveying line and an upper computer, the upper computer is in communication connection with the feeding conveying line, the hydraulic machine body, the finished product conveying line and the extrusion mechanism, and the feeding conveying line and the finished product conveying line are arranged on one side of the hydraulic machine body. An extrusion mechanism is arranged in the hydraulic machine body and comprises a lower die rotating mechanism, a pressure head switching mechanism and a stress base. Pressure is accurately compensated through a cold extrusion forming dynamic pressure compensation process, pressure compensation high-precision closed-loop control is achieved, and the self-adaptive capacity is enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cold extrusion forming, and particularly relates to a low-carbon steel cold extrusion forming production line and a forming dynamic pressure compensation process thereof. BACKGROUND

[0002] In the field of metal processing and manufacturing, the low-carbon steel cold extrusion forming process is widely used in the production of precision parts in the automobile, mechanical equipment and hardware product industries due to its advantages of improving the utilization rate of blank and strengthening the mechanical properties of parts. However, the current known low-carbon steel cold extrusion forming operation mode has many outstanding problems that restrict the development of the industry.

[0003] In the traditional production process, a large number of links from blank loading to finished product unloading after extrusion forming rely on manual operation. Workers need to frequently carry heavy low-carbon steel blanks and load and unload molds, which is extremely labor-intensive.

[0004] In the cold extrusion process, the uneven deformation of the low-carbon steel blank is prone to cause quality problems. The traditional constant blank holder force control cannot adapt to the stress requirements of different deformation stages of low-carbon steel, and the edge area is often cracked due to insufficient pressure or excessively deformed due to excessive pressure, causing fluctuations in the dimensional accuracy of the parts.

[0005] In terms of production efficiency, the speed and accuracy of manual operation are limited by physical strength and energy, and problems such as deviation of the placement position of the blank cause the equipment to idle and wait, the process connection is not smooth, and the production rhythm of the entire production line is difficult to improve, which cannot meet the demand of modern manufacturing for efficient production. SUMMARY

[0006] The present application relates to the technical field of cold extrusion forming, and particularly relates to a low-carbon steel cold extrusion forming production line and a forming dynamic pressure compensation process thereof.

[0007] In order to achieve the above-mentioned purpose, the technical scheme adopted by the present application is as follows: a low-carbon steel cold extrusion forming production line, comprising a feeding conveying line, a hydraulic machine main body composed of an upper beam, a movable beam and a lower beam, a hydraulic system, a main cylinder, an extrusion mechanism, a finished product conveying line and an upper computer, the upper computer being in communication connection with the feeding conveying line, the hydraulic machine main body, the finished product conveying line and the extrusion mechanism, characterized in that the feeding conveying line and the finished product conveying line are arranged on one side of the hydraulic machine main body, the extrusion mechanism is arranged in the hydraulic machine main body, and the extrusion mechanism comprises a lower die rotating mechanism, a pressure head switching mechanism and a stress base; The pressure head switching mechanism comprises a push oil cylinder, a backing plate, a sliding plate, a first pressure head and a second pressure head, the output end of the push oil cylinder is fixedly connected to one side of the sliding plate, the first pressure head and the second pressure head are fixedly connected to the lower plate surface of the sliding plate, the sliding plate slides on the lower plate surface of the backing plate, and the backing plate is fixedly connected below the movable beam; The lower mold rotating mechanism comprises a rotating mechanism and at least one lower mold, the rotating mechanism drives the lower mold to rotate, and the lower mold corresponds to the stressed base, the first pressing head or the second pressing head when being pressed; The pressing head switching mechanism is used to perform switching action when the lower mold rotates to the target station.

[0008] Further, the lower mold comprises a mold cavity, a mold frame and an ejection rod, and the mold cavity corresponds to the pressing head switching mechanism; The lower mold has four, namely the first lower mold, the second lower mold, the third lower mold and the fourth lower mold, and the first lower mold, the second lower mold, the third lower mold and the fourth lower mold are sequentially annular and can be pressed by the pressing head switching mechanism.

[0009] Further, a cold extrusion forming dynamic pressure compensation process is applied to the above-mentioned low-carbon steel cold extrusion forming production line, comprising the following process steps: S1, data acquisition and preprocessing; S2, generating a theoretical extrusion pressure curve; S3, monitoring the actual pressure and calculating the pressure deviation; S4, controlling the fuzzy neural network compensation model to output the compensation amount; S5, PID adaptive parameter dynamic adjustment; S6, hydraulic system dynamic response step S5; S7, compensation effect evaluation and iterative optimization.

[0010] Further, the data acquisition in S1 includes collecting physical parameters of the blank, including blank density ρ, blank hardness HV, blank yield strength σ s , real-time temperature T of the blank, geometric parameters of the lower mold, including extrusion ratio λ, die angle α, and device state parameters, including hydraulic system stiffness K s , pump flow Q; The S1 also includes establishing a constitutive relationship model of the blank Wherein, σ is the flow stress of the blank, σ0 is the strength coefficient of the blank, n is the strain hardening index of the blank, C is the strain rate sensitivity coefficient of the blank, ε is the equivalent strain of the blank, and ε´ is the equivalent strain rate of the blank.

[0011] Further, in the S2, the theoretical extrusion pressure P th is calculated based on a multi-field coupling model, and the calculation formula is Wherein, K1 is a three-dimensional pressure stress correction coefficient, σ s is the yield strength of the blank, λ is the extrusion ratio of the blank, K2 is the nonlinear index of the extrusion ratio, μ is the friction coefficient of the lower mold and the blank contact surface, α is the die angle of the lower mold, ΔT is the temperature change of the blank, and K3 is the temperature sensitivity coefficient.

[0012] Further, in the S3, the actual pressure P is collected in real time by the pressure sensor embedded in the first or second pressure head act , and the pressure deviation ΔP between the actual pressure and the theoretical extrusion pressure is calculated, and the calculation formula is: , wherein P th represents the theoretical extrusion pressure, and P act represents the actual pressure; the pressure deviation change rate calculation formula is: , wherein ΔP(t) represents the current time pressure deviation, ΔP(t-1) represents the previous time pressure deviation, and Δt represents the sampling time interval.

[0013] Further, in the S4, the pressure deviation ΔP, the pressure deviation change rate ΔP', the yield strength σ s of the blank, the extrusion ratio λ of the blank, and the real-time temperature T of the blank are taken as the input layer variables of the fuzzy neural network compensation model.

[0014] Further, in the S6, the hydraulic system receives the PID control signal of the step S5, adjusts the valve opening degree , and the calculation formula is: , wherein is the initial opening degree, K q is the valve gain, and ΔU is the compensation amount.

[0015] Further, in the S7, the compensation effect is judged based on the evaluation index, and if the requirement is not met, the model parameters or the control strategy are optimized reversely to form a closed loop evolution; The evaluation index includes the mean absolute percentage error MAPE evaluation and the dynamic performance index evaluation, Based on the actual pressure data collected in the S3 and the theoretical extrusion pressure curve in the S2, the mean absolute percentage error is calculated, and the calculation formula is: When the mean absolute percentage error MAPE exceeds the threshold value, the model parameter self-learning mechanism is triggered, and the blank constitutive model and the neural network weight are updated, wherein P thi is the theoretical extrusion pressure value of the i-th sampling point (from the theoretical extrusion pressure curve in the step S2); P acti is the actual extrusion pressure value of the i-th sampling point (from the pressure sensor collection in the step S3); i is the index of the sampling point (from 1 to N); N is the total number of sampling points participating in the calculation (i.e. the number of pressure data collected in the evaluation period).

[0016] Further, the threshold of the mean absolute percentage error MAPE is set to 3%, and when the mean absolute percentage error MAPE is greater than 3%, the model parameter self-learning mechanism updates the calculation formula of the neural network weight value as follows: wherein, represents the current weight value, η represents a learning rate, δ j represents an error term, represents an input layer variable, α represents a learning rate, represents the weight value of the connection between the ''input layer i-th neuron'' and ''hidden layer j-th neuron'' in the neural network at the t+1 iteration.

[0017] When the dynamic performance index is unqualified, the PID parameter is preferentially optimized.

[0018] Compared with the prior art, the low carbon steel cold extrusion forming production line and the forming dynamic pressure compensation process have the advantages that: 1. High degree of automation, material transportation is achieved through feeding conveying line and finished product conveying line, four-station cycle pressing, and the pressure head switching mechanism can randomly switch the pressure head, saving labor cost; 2. High-precision closed-loop control is achieved, the mean absolute percentage error MAPE threshold is accurately bottomed, and the pressure control precision is locked within 3%, solving the problem of pressure deviation caused by billet fluctuation and die wear; 3. Enhanced self-adaptive capability, according to the real-time collected multi-dimensional parameters such as pressure deviation ΔP, deviation change rate ΔP', billet yield strength σ s , extrusion ratio λ and billet temperature T, dynamic optimization of fuzzy rule weight and PID control parameter is achieved, adaptive control in the cold extrusion process is realized, and different specifications of billets can be adapted without frequent manual intervention; 4. Strong practicability: modular design is adopted, the pressure sensor, electro-hydraulic servo valve and other hardware devices of the existing production line can be directly reused, intelligent transformation is realized through algorithm upgrading, and the transformation cost is reduced; meanwhile, an interactive interface is arranged, manual intervention and rule import are supported, the algorithm autonomy and process flexibility are considered, and the application in the existing cold extrusion production line is facilitated. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is a whole schematic view of the present application; Figure 2 is a side view schematic view of the present application; Figure 3Figure 2 is a perspective view of the lower mold rotating mechanism of the present application; Figure 4 Figure 3 is a top view of the lower mold rotating mechanism of the present application; Figure 3 Figure 5 Figure 4 is a side view of the lower mold rotating mechanism of the present application; Figure 6 Figure 5 is a sectional view of the lower mold rotating mechanism of the present application; Figure 7 Figure 6 is a schematic view of the gear driving mechanism of the present application; Figure 8 Figure 7 is a schematic view of the pressing head switching mechanism of the present application; Figure 9 Figure 8 is a schematic view of the process flow of the present application; Figure 10 Figure 9 is a schematic view of the frame flow of the present application; Figure 11 Figure 10 is a schematic view of the frame flow of step S7; In the figure: 2, upper computer, 3, feeding conveying line, 4, feeding manipulator, 5, hydraulic machine main body, 6, discharging manipulator, 7, finished product conveying line, 9, lower mold rotating mechanism, 10, pressing head switching mechanism, 11, force receiving base, 12, rotating platform, 13, gear driving mechanism, 131, motor, 132, driving gear, 133, driving push block, 134, push block cylinder, 135, driven gear, 136, driven push block, 137, connecting gear, 138, base, 14, rotary gear disc, 15, first lower mold, 151, mold cavity, 152, ejector rod, 153, cushion block, 16, second lower mold, 17, third lower mold, 18, fourth lower mold, 19, pushing oil cylinder, 20, cushion plate, 21, sliding plate, 22, first pressing head, 23, second pressing head. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0021] ​In the description of the present application, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application., and the terms "first", "second" and the like are only for distinguishing the names of various parts and have no primary and secondary relationship, so they cannot be understood as limiting the present application.

[0022] Embodiment 1, the present application provides a low carbon steel cold extrusion forming production line, please see the description of the drawings Figures 1-8 As shown in the description of the drawings, it comprises a feeding conveying line 3, a hydraulic machine main body 5 composed of an upper beam, a movable beam and a lower beam, a hydraulic system, a main cylinder, an extrusion mechanism, a finished product conveying line 7 and an upper computer 2, the upper computer 2 is in communication connection with the feeding conveying line 3, the hydraulic machine main body 5, the finished product conveying line 7 and the extrusion mechanism, the feeding conveying line 3 and the finished product conveying line 7 are arranged on one side of the hydraulic machine main body 5, the extrusion mechanism is arranged in the hydraulic machine main body 5, the extrusion mechanism comprises a lower die rotating mechanism 9, a pressure head switching mechanism 10 and a force receiving base 11; the force receiving base 11 is fixedly installed on the lower beam of the hydraulic machine; a knockout cylinder is further arranged below the movable beam, which is used to lift the movable beam away from the lower die rotating mechanism 9.

[0023] The pressure head switching mechanism 10 comprises a push oil cylinder 19, a backing plate 20, a sliding plate 21, a first pressure head 22 and a second pressure head 23, the output end of the push oil cylinder 19 is fixedly connected to one side of the sliding plate 21, the lower plate surface of the sliding plate 21 is fixedly connected to the first pressure head 22 and the second pressure head 23, the sliding plate 21 slides on the lower plate surface of the backing plate 20, and the backing plate 20 is fixedly connected below the movable beam; The lower die rotating mechanism 9 comprises a rotating mechanism and at least one lower die, the rotating mechanism drives the lower die to rotate, and the lower die corresponds to the force receiving base 11, the first pressure head 22 or the second pressure head 23 when being pressed; The pressure head switching mechanism 10 is used to perform switching action when the lower die rotates to the target station.

[0024] The first pressure head 22 and the second pressure head 23 are used alternately, which enhances the service life of the pressure head. Different pressure heads can also be replaced to correspond to different lower dies for pressing operation. The target station is the pressing station inside the hydraulic machine main body 5, when the first pressure head 22 and the second pressure head 23 are to be switched, the push oil cylinder 19 pushes the sliding plate 21 forward or backward, which can switch the first pressure head 22 and the second pressure head 23, so that the first pressure head 22 and the second pressure head 23 correspond to the lower die for pressing operation.

[0025] The lower mold comprises a mold cavity 151 corresponding to the pressure head switching mechanism 10, an ejection rod 152, and a mold frame 154. The ejection rod 152 is arranged in the mold frame 154 and corresponds to the mold cavity 151. A cushion block 153 is further arranged on the periphery of the ejection rod. The cushion block 153 is arranged between the mold frame 154 and the mold cavity 151 to enhance the stability of the mold cavity 151.

[0026] The lower mold has four structures, as shown in the figure, which are the first lower mold 15, the second lower mold 16, the third lower mold 17, and the fourth lower mold 18. The first lower mold 15, the second lower mold 16, the third lower mold 17, and the fourth lower mold 18 are sequentially rotated and correspondingly pressed by the pressure head switching mechanism 10. Specifically, when pressing, for example, the first lower mold 15 is being pressed, the second lower mold 16 is rotated to the demolding mechanism, and at the same time, the fourth lower mold 18 is rotated to the side of the feeding mechanical arm 4 for feeding. Figure 4 On one side of the stressed base 11, a demolding mechanism is also fixedly arranged. The demolding mechanism comprises a demolding cylinder. The extension end of the demolding cylinder is fixedly connected with the bottom of the ejection rod 152. The demolding mechanism demolds the finished products in the first lower mold 15, the second lower mold 16, the third lower mold 17, and the fourth lower mold 18, respectively.

[0027] In this embodiment, the rotating mechanism comprises a rotating platform 12 and a gear driving mechanism 13, as shown in the figure, which belongs to the prior art. The rotating platform 12 is uniformly distributed with a plurality of workbenches for placing lower molds on the periphery. The rotating platform 12 is provided with a rotating gear plate 14 with a gear at the lower part. The rotating gear plate 14 is engaged with the gear driving mechanism 13. The rotating platform 12 rotates cyclically under the driving of the gear driving mechanism 13.

[0028] Figure 7 The gear driving mechanism 13 comprises a motor 131, a speed reducer, a base 138, and a gear set. The gear set is arranged on the base 138 and comprises a driving gear 132, a driven gear 135, and a connecting gear 137. The motor 131 is connected with the speed reducer. The driving shaft of the speed reducer is connected with the driving gear 132. The driving gear 132 is engaged with the driven gear 135. The driven gear 135 is engaged with the connecting gear 137.

[0029] The top end of the driving gear 132 is fixedly connected with a driving push block 133. The two ends of the driving push block 133 are fixedly provided with a push block cylinder 134. The top end of the driven gear 135 is fixedly provided with a driven push block 136.

[0030] The driven gear 135 is an intermittent gear. The intermittent surface of the intermittent gear is consistent with the extension direction of the driven push block 136. The height of the intermittent surface of the intermittent gear is equal to the height of the driving gear 132.

[0031] The driven gear 135 is an intermittent gear. The intermittent surface of the intermittent gear is consistent with the extension direction of the driven push block 136. The height of the intermittent surface of the intermittent gear is equal to the height of the driving gear 132. ​

[0032] The intermittent time of the driven gear 135 is the time for the movable beam to press the blank in the lower mold.

[0033] The driven push block 136 is in the shape of a "racket", the arc surface of the "racket" handle matches the outer arc surface of the push block cylinder 134, and the distance between the end faces of the two push block cylinders 134 is greater than the maximum diameter of the driven push block 136.

[0034] When the motor 131 drives the driving gear 132 to rotate, the driving push block 133 will also rotate, and the driven push block 136 will also be driven to rotate by the driven gear 135. Since the direction of the driven push block 136 is the same as the direction of the intermittent surface of the intermittent gear, when the driven push block 136 rotates between the two push block cylinders 134, the intermittent surface of the driven gear 135 cannot be meshed with the gear, and the driven gear 135 will stop rotating, and the connecting gear 137 will also stop rotating. When the push block cylinder 134 on the left side is rotated to the arc surface at the root of the driven push block 136 under the drive of the driving gear 132, the driven push block 136 will be pushed by the push block cylinder 134. When the two push block cylinders 134 on the driving gear 132 continue to rotate, the driven push block 136 will drive the driven gear 135 to rotate, so that the driven gear 135 is meshed with the driving gear 132 again, and drives the connecting gear 137 to rotate, and the connecting gear 137 drives the rotary gear disc 14 to rotate. The two push block cylinders 134 on the driving gear 132 alternately push the driven push block 136, so that the work is continuously and cyclically performed.

[0035] In the embodiment, the feeding conveying line 3 and the finished product conveying line 7 are both chain plate conveying lines, and the feeding and discharging can be performed by robots to save labor cost.

[0036] The production line implements the following steps: starting the rotating mechanism, the first lower mold 15, the second lower mold 16, the third lower mold 17 and the fourth lower mold 18 are driven to rotate, and the feeding conveying line 3 and the finished product conveying line 7 are driven to rotate. Figure 2 、 Figure 4As shown, the feeding manipulator 4 feeds the blank into the first lower mold 15, the rotating mechanism drives the first lower mold 15 to rotate to the pressing station inside the hydraulic machine main body 5, the fourth lower mold 18 rotates to the side of the feeding manipulator 4, the feeding manipulator 4 feeds the fourth lower mold 18, at the same time, the first lower mold 15 is above the force receiving base 11, the main cylinder drives the movable beam to descend, the movable beam drives the first pressing head 22 to press, after the pressing is completed, the ejecting cylinder under the movable beam lifts the movable beam, the rotating mechanism rotates again, the first lower mold 15 rotates clockwise to the position of the second lower mold 16, at the same time, the fourth lower mold 18 rotates to the pressing station inside the hydraulic machine main body 5 to perform the pressing operation, the rotating mechanism rotates again, the first lower mold 15 rotates to the position of the third lower mold 17, the position of the third lower mold 17 corresponds to the demolding mechanism, the finished product in the first lower mold 15 is above the demolding mechanism, the demolding cylinder ejects to demold, the discharging manipulator 6 takes out the finished product and places it on the finished product conveying line 7, the rotating mechanism rotates again, the first lower mold 15 rotates to the side of the feeding manipulator, the feeding manipulator 4 puts the blank into the first lower mold 15, at the same time, the fourth lower mold 18 rotates to above the demolding mechanism, after demolding, the discharging manipulator 6 takes away the finished product in the fourth lower mold 18 and places it on the finished product conveying line 7, the rotating mechanism rotates again to bring the first lower mold 15 into the hydraulic machine main body 5 to perform the pressing, and the same is repeated until all the products are pressed.

[0037] As shown in FIG. 2, a cold extrusion forming dynamic pressure compensation process is applied to the low-carbon steel cold extrusion forming production line. Figure 9 and Figure 10 As shown in FIG. 2, a cold extrusion forming dynamic pressure compensation process is applied to the low-carbon steel cold extrusion forming production line. In order to make the low-carbon steel cold extrusion forming effect better, a host computer 2 and a PLC controller are arranged, the host computer cooperates with the PLC controller to control the compensation process, the PLC controller is built-in with a PID algorithm module, the PLC controller includes an input module, a processing module and an output module, and specifically includes the following process steps: S1, data acquisition and preprocessing; S2, generating a theoretical extrusion pressure curve; S3, monitoring the actual pressure and calculating the pressure deviation; S4, controlling the fuzzy neural network model to output the compensation amount; S5, PID adaptive parameter dynamic adjustment; S6, hydraulic system dynamic response step S5; S7, compensation effect evaluation and iterative optimization.

[0038] In step S1, data acquisition includes collecting physical parameters of the blank, including the density p of the blank, the hardness HV of the blank, the yield strength s s, the physical parameters of the blank belong to the inherent properties of the blank, and the real-time temperature T of the blank can be measured by an infrared temperature measuring instrument installed on the feeding conveying line. The lower die geometric parameters are collected, including the extrusion ratio λ, the die angle α, and the equipment state parameters, the lower die geometric parameters are the fixed design parameters of the lower die; and the hydraulic system stiffness K s , pump flow Q; the hydraulic system stiffness K s can be calculated based on the cylinder diameter, piston rod diameter and oil volume elastic modulus, etc., the pump flow Q is obtained by being installed on the output oil way or main oil way of the main pump, and these basic data are pre-set in the upper computer. The upper computer 2 further issues to the PLC controller.

[0039] In this embodiment, the extrusion ratio λ of the Q235 low carbon steel shaft part (initial diameter ϕ40mm, target finished product diameter ϕ20mm) can be calculated, the extrusion ratio is defined as the ratio of the initial cross-sectional area A0 of the blank to the finished product cross-sectional area A1, and the calculation formula is: , The forward cold extrusion process is adopted, and the lower die is formed by a 5000T hydraulic machine, the lower die is a conical structure, and the lubrication mode is MoS2 dry film lubrication.

[0040] The density ρ of the blank, the hardness HV of the blank, and the yield strength σ s of the blank can be directly read by the characteristics of the blank, the geometric parameters of the die are determined by the design drawing, for example, the die angle α of the lower die is 45°, the hydraulic system stiffness K s , and the pump flow Q are tested by the hydraulic system (step pressure method), K s =1.2×10 9 N / m, for the dynamic response PID algorithm module in step S6, the rated pump flow Q of the hydraulic pump is 60L / min=1×10 −3 m 3 / s, an infrared temperature measuring instrument is installed on the side of the lower die to detect the real-time change of the blank temperature, after the data collection, the blank constitutive relation model is established , wherein σ is the flow stress of the blank, σ0 is the strength coefficient of the blank, n is the strain hardening index of the blank, C is the strain rate sensitivity coefficient of the blank, and ε is the equivalent strain of the blank.

[0041] We set the strength coefficient σ0=500MPa, the strain hardening index n=0.15, and the strain rate sensitivity coefficient C=0.03.

[0042] The equivalent strain ε, in cold extrusion, the equivalent strain and the extrusion ratio approximately satisfy ε=lnλ, so: ε=ln4≈1.386, Assuming the extrusion speed v = 10 mm / s, the blank thickness h = 5 mm at the die inlet, then: , Finally, these data are substituted into the blank constitutive relationship model: .

[0043] The method for obtaining the blank constant in the blank constitutive relationship model includes: determining the stress-strain curve of the low-carbon steel in the strain rate range of 10 −3 ~10 2 s −1 by the uniaxial tensile test, and fitting the strength coefficient σ0, the strain hardening index n and the strain rate sensitive coefficient C of the blank by the nonlinear least square method, wherein the value range of σ0 is 300-600 MPa, n is 0.1-0.2, and C is 0.01-0.05.

[0044] S2, generating a theoretical extrusion pressure curve, first calculating the theoretical extrusion pressure P th based on the multi-field coupling model, and the calculation formula is: , wherein K1 is a three-dimensional pressure stress correction coefficient, σ s is the yield strength of the blank, λ is the extrusion ratio of the blank, K2 is the nonlinear index of the extrusion ratio, μ is the friction coefficient of the lower die and the contact surface of the blank, α is the die angle of the lower die, ΔT is the temperature change amount of the blank, and K3 is the temperature sensitive coefficient.

[0045] In cold extrusion, the blank is in a three-dimensional pressure stress state (extruded from all around by the lower die), and the yield strength σ s is measured by the uniaxial tensile test (unidirectional stress). The three-dimensional pressure stress will make the real deformation resistance of the blank much greater than σ s , and K1 is used to quantify this “stress state strengthening effect”.

[0046] By calculating the flow stress σ of the blank, the correction coefficient of the theoretical pressure formula can be calibrated, and σ is the “calibration source” of the correction coefficient. By comparing σ s , under three-dimensional pressure stress, the flow stress σ ≈ 541.5 MPa ≈ 2.3σ s (σ s ≈235 MPa), assuming K2=0.7 first, and then substituting it into the theoretical extrusion pressure formula to verify it inversely, it is known that P th ≈770 MPa (theoretical extrusion pressure of cold extrusion), λ=4, μ=0.1, α=45°, (cotα=1), ΔT=30°, K3=−0.002, and K1=1.2 is calculated by the formula , when K1=1.2, K1*σs =282MPa, close to σ / 2 (541.5 / 2≈270.75), which means: K1 only corrects part of the strengthening of three-directional stress (the remaining strengthening is covered by the correction of K2 strain hardening); K1 tube stress state, K2 tube strain hardening, K3 tube temperature, this "step-by-step correction" balances the complexity and accuracy of the formula. Calibration of process correction coefficients K1, K2, K3: depends on the calculation of the blank constitutive relation model σ.

[0047] The cold extrusion process is affected by many factors such as three-directional compression stress, strain hardening, temperature change, friction, etc. If a single coefficient is used to correct all factors, the formula will be too complex to be used in engineering; step-by-step correction (K1 tube stress state, K2 tube strain hardening, K3 tube temperature) not only simplifies the formula, but also covers the main influencing factors through the division of each coefficient.

[0048] Then K1, K2, K3 are brought into the formula That is, the theoretical extrusion pressure can be calculated.

[0049] In the actual production of cold extrusion, due to the variety of working conditions (such as changing the blank from Q235 to 304 stainless steel, adjusting the extrusion ratio from 4 to 5, changing the die temperature, etc.): if only rely on "simulation test", each time the working condition changes, it needs to be tested again (takes several hours or even longer); through the calibrated theoretical formula, only by replacing the blank yield strength, extrusion ratio of the blank and other parameters, the theoretical extrusion pressure of the corresponding working condition can be calculated in seconds, which greatly improves the engineering efficiency. Let the theoretical extrusion pressure calculation formula have multi-condition universality.

[0050] In the cold extrusion process, the required pressure is dynamically changing, the main reasons are as follows: Initial stage: the blank just contacts the lower die, plastic deformation begins, and the deformation resistance starts from zero and rises rapidly.

[0051] Stable extrusion stage: the blank continues to pass through the lower die, the deformation degree (the equivalent strain ε of the blank) reaches the maximum value and remains stable, but due to the strain hardening effect of the blank, the flow stress still increases slowly, resulting in a slow increase in the required pressure.

[0052] End stage: the remaining length of the blank is very short, the deformation begins to expand to the entire remaining blank, the shape of the deformation zone changes, resulting in fluctuating changes in the pressure.

[0053] Change of friction condition: at the beginning of extrusion, the MoS2 dry film may not have spread evenly, and the friction is larger.

[0054] As the extrusion proceeds, the temperature rise of the lower die may cause changes in the performance of the MoS2 dry film, and the friction also changes.

[0055] Variation of contact area: for forward extrusion, as the first or second ram 22 or 23 advances, the contact area of the billet with the lower die also changes, which directly affects the total friction.

[0056] Therefore, in the extrusion process, the pressure is dynamically changing. The host computer 2 uses the input billet parameters, lower die parameters and the preset constitutive relationship model σ to calculate and predict the theoretical extrusion pressure curve that changes with time or displacement required to complete the entire extrusion process under the current conditions, i.e. a continuous theoretical extrusion pressure curve. The specific process is as follows: 1. Calculate the instantaneous strength (flow stress σ) of the billet in the dynamic deformation process in real time through the billet constitutive relationship model, and the calculation formula is: .

[0057] 2. Establish the relationship with displacement / time: Equivalent strain ε: In the extrusion process, the equivalent strain ε has a direct mathematical relationship with the instantaneous billet length or master cylinder displacement S (ε increases from 0 to lnλ). Therefore, ε is a function of displacement S: ε = f(S) Equivalent strain rate ε': related to the extrusion speed v. In the case of constant extrusion speed v, ε' can be expressed as a function of master cylinder displacement S.

[0058] 3. Iterative calculation to generate the curve: The host computer discretizes the entire extrusion stroke (from displacement S = 0 to S = maximum value) into thousands of points.

[0059] For each displacement point S i , the process of generating the theoretical extrusion pressure curve is as follows: a. Calculate the strain ε i and strain rate ε' i corresponding to this position; b. Substitute ε i and ε' i into the billet constitutive relationship model to calculate the instantaneous flow stress σ i ; c. Substitute σ i into the theoretical pressure formula, i.e. , Thus, the theoretical extrusion pressure value corresponding to the displacement point S i is obtained.

[0060] Connecting all displacement points S i and the calculated theoretical extrusion pressure values forms the final theoretical extrusion pressure curve.

[0061] The curve is then sent to the PLC controller as a reference for real-time pressure monitoring and deviation calculation in S3.

[0062] In S3, the actual pressure P is collected in real time by the pressure sensor embedded in the first or second pressure head 22 or 23 act , and the pressure deviation ΔP between the actual pressure and the theoretical extrusion pressure is calculated, with the formula being: , wherein P th represents the theoretical extrusion pressure value, and P act represents the actual pressure value; and the pressure deviation change rate is calculated with the formula being: , wherein ΔP(t) represents the current pressure deviation, ΔP(t-1) represents the previous pressure deviation, and Δt represents the sampling time interval.

[0063] In this embodiment, the pressure sensor is a strain pressure sensor installed between the contact surface of the first or second pressure head 22 or 23 and the die cavity 151. The strain signal is converted to voltage by a charge amplifier, and the voltage signal converted by the charge amplifier is received by the input module of the PLC controller. According to the calibration curve of the pressure sensor (which is the corresponding relationship between the voltage value and the pressure value determined after the pressure sensor is factory-calibrated or field-calibrated), the voltage signal is converted to the actual pressure value P act . The input module transmits the actual pressure value P act to the processing module in the PLC controller, which calculates the pressure deviation ΔP and the pressure deviation change rate ΔP' in real time as the basis for compensation decision, and the output module uploads the pressure deviation ΔP and the pressure deviation change rate ΔP' to the upper computer 2 upon receiving the signal from the processing module.

[0064] In S4, the fuzzy neural network compensation model outputs a compensation amount, and the upper computer 2 takes the pressure deviation ΔP, the pressure deviation change rate ΔP', the yield strength σ s of the blank, the extrusion ratio λ of the blank, and the real-time temperature T of the blank as the input layer variables of the fuzzy neural network compensation model, and the fuzzy neural network compensation model is run through the fuzzy neural network compensation controller.

[0065] In this embodiment, the fuzzy neural network compensation controller is communicatively connected to the upper computer 2 through an Ethernet interface, and the fuzzy neural network compensation model includes an input layer, a data preprocessing module, a fuzzy reasoning module, and a rule base management module. The input layer receives the pressure deviation ΔP reflecting the deviation degree of the actual extrusion pressure from the theoretical pressure value, the pressure deviation change rate ΔP' reflecting the fluctuation trend and intensity of the pressure, and the yield strength σ s, extrusion ratio λ reflecting the degree of deformation of the blank, and physical parameters reflecting the real-time temperature T of the blank in the cold extrusion process. The data preprocessing module normalizes each input variable and maps it to the interval [-1, 1].

[0066] For example, the pressure deviation ΔP ranges from [-50, 50] MPa, and after normalization , The real-time temperature T of the blank ranges from [20, 150] °C, and after normalization T' is the normalized temperature value.

[0067] The fuzzy inference module maps the input vector preprocessed by the data preprocessing module to the control compensation amount, including: 1. Fuzzy layer: converts the precise values in the data preprocessing module to fuzzy set membership degrees, i.e., fuzzy results; 2. Rule inference layer: executes fuzzy rules based on the data in the rule base management module to obtain fuzzy conclusions; 3. De-fuzzy layer: converts the fuzzy conclusions to precise control amounts, and through the barycentric method (weighted average method), the "fuzzy conclusions" of multiple active rules are fused into precise compensation amounts ΔU.

[0068] The rule base management module includes a fuzzy rule base for matching and comparing with fuzzy set membership degrees. The fuzzy rule base adopts a dual-redundancy storage architecture (RAM real-time call + Flash persistent backup) to store multiple core fuzzy rules. Each fuzzy rule contains unique ID, premise condition membership parameters, conclusion membership parameters, weight (0~1), cumulative activation number, average error contribution, etc. High-efficiency indexing is supported, and a binary tree index is established according to the input variable priority (pressure deviation ΔP, pressure deviation change rate ΔP', blank yield strength σ s , extrusion ratio λ, and real-time temperature T of the blank in the cold extrusion process). High-frequency trigger rules (top 20) are also pre-stored in the fast query table to ensure that the rule call delay is ≤50 μs, meeting the real-time requirements of 100 Hz sampling frequency.

[0069] The rule base management module filters out effective fuzzy rules (rules that match the current fuzzy results and have qualified error contribution) from the fuzzy rule base. The fuzzy inference module executes rule inference based on the fuzzy results and effective fuzzy rules to obtain fuzzy conclusions. Then, the fuzzy inference module uses the barycentric method de-fuzzy algorithm to convert the fuzzy conclusions to precise compensation amounts ΔU. Finally, the fuzzy inference module outputs the control compensation amount ΔU to the upper computer 2.

[0070] S5. Dynamic adjustment of PID adaptive parameters; the host computer transmits the precise compensation amount ΔU to the PLC controller, and the PLC's built-in PID algorithm module adjusts the PID parameters according to the compensation amount ΔU, dynamically adjusting the PID parameter ratio (K). p ), integral (K) i ), differential (K) d The parameters are used to generate the control signal for the electro-hydraulic servo valve. For example, the calculated precise compensation amount ΔU = +12MPa, ΔU > 10MPa, is then used to calculate the output current signal I. I0 is the intermediate value, K I This is for calibrating the gain.

[0071] S6, Hydraulic system dynamic response step S5; Hydraulic system dynamic response PLC controller, electro-hydraulic servo valve receives control signal from PLC controller and adjusts valve opening degree. This drives the master cylinder pressure to rapidly approach the theoretical extrusion pressure value. The calculation formula is as follows: ,in, For the initial aperture, K q Let ΔU be the valve gain, and ΔU be the compensation amount. In this embodiment, let ΔU be the valve gain. K q Given 2%, ΔU = +2MPa, substitute these values ​​into the formula: , As the opening of the electro-hydraulic servo valve increases from 30% to 34%, the hydraulic pump flow increases, and the main cylinder pressure rises rapidly, approaching the theoretical extrusion pressure value.

[0072] Assuming ΔU = -5MPa, calculate The opening of the electro-hydraulic servo valve is reduced from 30% to 20%, the hydraulic pump flow rate is reduced, and the master cylinder pressure drops (to eliminate overpressure).

[0073] The variable pressure sensor collects P every 10ms act The new pressure deviation and deviation change rate are calculated and used as the control input for the next round.

[0074] In the S7, as Figure 11 As shown, the compensation effect evaluation and iterative optimization includes judging the compensation effect based on the evaluation index. If the requirements are not met, the model parameters (bill constitutive relationship model and fuzzy neural network compensation model) or control strategy are optimized in reverse to form a closed-loop evolution. The compensation effect evaluation and iteration are run through the rule base management module, which also includes a rule evaluation module, a rule optimization submodule, a rule update submodule, a rule conflict resolution submodule, and an interaction submodule.

[0075] The rule evaluation module assesses whether each fuzzy rule helps reduce or increases deviation during the control process. It can quantify the effectiveness of fuzzy rules in real time and calculate the error contribution (EC) of each fuzzy rule. i The mean absolute percentage error (MAPE) is calculated to evaluate the fit between the theoretical extrusion pressure curve and the actual pressure curve, and to determine the accuracy of the billet constitutive model. For example, for every 10 products produced, the activation frequency (as a percentage of the total number of triggers) and error contribution EC of each fuzzy rule are calculated. i (The average pressure deviation when the fuzzy rule is triggered) is used to generate a validity score of 0-100 (a score of 80 or above is considered a "valid rule"). In this embodiment, the error contribution EC i The threshold is 8 MPa, and the mean absolute percentage error (MAPE) threshold is set to 3%.

[0076] The rule optimization submodule receives diagnostic results from the evaluation module and performs fine-tuning and optimization on poorly performing fuzzy rule sets. For example, it can streamline the fuzzy rule library, delete rules with validity scores <60, merge similar fuzzy rules (e.g., those with precondition vector similarity >85%) using a cosine similarity algorithm (threshold 0.85), and retain redundant fuzzy rules after weighted averaging. Another function is weight redistribution: normalizing the weights of remaining fuzzy rules according to their validity scores to ensure high error contribution coefficients (ECs). i The weight of rules (such as those with an error of <5MPa) is increased by 10% to 20%. If the performance of the hydraulic system still fails to meet the standards after optimization, a new fuzzy rule generation process is triggered.

[0077] The rule update submodule is used to update and upgrade the billet constitutive model and the fuzzy neural network compensation model.

[0078] For example: Automatic updates are triggered when the following conditions are detected: pressure fluctuation standard deviation > 8 MPa for 3 consecutive batches of products; billet yield strength σs deviation > 15% from historical value (e.g., when changing material grade); manual triggering via host computer (e.g., when changing mold). Incremental learning: Based on 500 sets of sample data (input variables + optimal compensation amount) for new conditions, mapping relationships are extracted through CART decision tree to generate 8-12 new rules. After simulation verification (error < 5 MPa), these rules are added to the fuzzy rule base.

[0079] The rule conflict resolution submodule is responsible for arbitration and ensuring that a unified and reasonable control command is output when multiple fuzzy rules are activated at the same time and their output conclusions contradict each other.

[0080] For example, the rule conflict resolution sub-module in real-time monitoring of the reasoning process, if the same input vector triggered opposite conclusions of the rules (such as ΔU=PB "substantially increased" and ΔU=NB "substantially reduced"), calculate the conflict intensity wherein, represent the same input, all triggered fuzzy rules in the "most aggressive compensation direction", represent the same input, all triggered fuzzy rules in the "most conservative compensation direction", and the greater the gap, the more intense the contradiction between the rules. The calculation of the conflict intensity of the "implicit contradictions" of fuzzy rules into "explicit numerical value", support intelligent conflict processing of cold extrusion pressure control.

[0081] For example, in the cold extrusion process, the temperature of the blank rises sharply (T=120°C, much higher than normal), while the pressure deviation ΔP=+30MPa (the actual pressure value is higher than the theoretical extrusion pressure value).

[0082] Fuzzy rule triggering and conflict: Rule 1 and Rule 2 are two specific fuzzy rules stored in the fuzzy rule library.

[0083] Fuzzy rule 1 (high temperature + positive deviation): conclusion (requires "substantial pressure reduction"); Fuzzy rule 2 (new lower die not updated rule): conclusion (error requires "pressure increase"); Calculate the conflict intensity: C=|−20−15|=35MPa, (more than 20% full scale, trigger high conflict strategy).

[0084] When the error contribution EC2 of fuzzy rule 2 is detected =25MPa, (much higher than the threshold value 8MPa), temporarily disable fuzzy rule 2; only execute the conclusion of fuzzy rule 1 (ΔU min =−20MPa), the pressure quickly falls to the theoretical extrusion pressure value; After production, the rule update submodule automatically learns new conditions and corrects the error of fuzzy rule 2.

[0085] Resolution strategy: If the conflict intensity C≤20% full scale (such as ≤20MPa): take the weighted average compensation amount; If the conflict intensity C> 20% full scale: temporarily disable the low effectiveness rule and enable the backup fuzzy rule (based on the historical optimal solution); Three consecutive conflicts: trigger the rule update submodule to relearn.

[0086] The interaction sub-module is used to send alarm information to the host computer 2, and pop up the human-computer interaction interface to request the engineer to manually adjust the fuzzy rule or PID parameter. The HMI interface of the host computer 2 can support manual creation / import of fuzzy rules, which directly act on the control logic. The XML format rule package import / export is supported, which is convenient for copying the mature process rule library. Rules can be added / removed, and operation audit logs are recorded.

[0087] For example, when a new lower die is put into production by cold extrusion, the blank is a custom alloy (the yield strength of the blank σ s fluctuates by 20%), and the algorithm initial rule control fluctuates by ±15 MPa; Through the HMI interface of the host computer 2, manually add rules: IF (ΔP = positive deviation, T = high temperature, λ = large deformation), THEN (ΔU = large pressure reduction, weight = 0.8); After the rule is imported, the pressure fluctuation is immediately reduced to ±5 MPa, shortening the debugging cycle.

[0088] The evaluation index simultaneously performs mean absolute percentage error (MAPE) evaluation, fuzzy rule error contribution degree evaluation, and dynamic performance index evaluation.

[0089] In this embodiment, the threshold value of the mean absolute percentage error (MAPE) is set to 3%, and when the MAPE is greater than 3%, the calculation formula of the neural network weight value of the self-learning mechanism is updated as follows: wherein, W represents the current weight value, η represents the learning rate, δ j represents the error term, X represents the input layer variable, α represents the learning rate, W represents the weight value of the connection between the “i-th neuron in the input layer” and the “j-th neuron in the hidden layer” in the neural network at the t+1 iteration.

[0090] The mean absolute percentage error (MAPE) is calculated according to the following formula: When the mean absolute percentage error (MAPE) exceeds the threshold value, the model parameter self-learning mechanism is triggered, and the blank constitutive model and the neural network weight value are updated, wherein: P thi is the theoretical extrusion pressure value of the i-th sampling point (from the theoretical extrusion pressure curve of step S2); P actiPi is the actual extrusion pressure value of the i th sampling point (pressure sensor acquisition from step S3); i is the index of the sampling point (from 1 to N); N is the total number of sampling points participating in the calculation (i.e. the number of pressure data collected within the evaluation period).

[0091] When the mean absolute percentage error MAPE exceeds the threshold value, the model parameter self-learning mechanism is triggered, the model includes the blank constitutive model and the fuzzy neural network compensation model, the fuzzy neural network compensation model further includes a neural network weight layer, and self-learning is realized through neural network weight adjustment to optimize the fuzzy rule weight or compensation accuracy, and the blank constitutive model and the neural network weight are updated.

[0092] Because in the cold extrusion process, the extrusion process is dynamic and continuous (the pressure changes with time or stroke), pressure data at multiple time points need to be collected (such as collecting 1 time every 0.1 seconds, and N = 100 for 10 seconds); N determines the statistical range of the mean absolute percentage error MAPE: the larger N is, the more complete the process stage covered is, and the mean absolute percentage error MAPE can better reflect the overall actual pressure control accuracy.

[0093] Suppose that in the cold extrusion process, the theoretical extrusion pressure of 5 points is: (theoretical pressure curve varying with stroke); The pressure deviation is small, and the following is calculated: , 2.49%<3% (threshold value), the pressure control accuracy meets the standard, and the self-learning mechanism is not triggered.

[0094] If , (the pressure deviation is too large in a certain stage, and the hardness of the simulated blank suddenly changes), , 3.33%>3% (threshold value), the pressure control accuracy is insufficient, and the self-learning mechanism is triggered to update the blank constitutive relationship model and the neural network weight.

[0095] After the dynamic performance index evaluation is stable, the following conditions need to be met: | ΔP | ≤ 3MPa (qualified), response time ≤ 0.5s, and overshoot ≤ 5% (qualified), then it is qualified.

[0096] Error contribution degree EC i ≤ 8MPa is qualified, and the error contribution degree EC iThe influence of a single fuzzy rule on the pressure deviation (unit: MPa) is represented. The influence of a single fuzzy rule is determined by its weight in the compensation decision. When multiple fuzzy rules are triggered, the fuzzy rule with a higher weight contributes more to the final compensation.

[0097] In the verified strategy, each control strategy (a set of PID parameters and fuzzy rule weights) is verified to be qualified, and the fuzzy neural network compensation model saves these configurations. These configurations are used for subsequent compensation of similar working conditions, realizing strategy reuse and optimization.

[0098] If the error contribution EC i is not qualified, the fuzzy rule optimization or new fuzzy rule generation is triggered. This is because when the overall performance is qualified but the individual fuzzy rule performs poorly, the problem lies in the "knowledge base" of intelligent compensation itself, which needs to be finely adjusted. When the dynamic performance index is not qualified, the PID parameters are optimized first. This is because the PID is the basic controller, and its adjustment can usually quickly and effectively solve most general control problems.

[0099] If the optimization is still not qualified for three consecutive times, the host computer 2 alarm is triggered, and the rules or PID parameters can be manually adjusted.

[0100] The billet constitutive model update can adopt machine learning methods (such as regression analysis, Gaussian process model, etc.) to dynamically correct the parameters of the billet constitutive model.

[0101] The neural network weight update can adopt reinforcement learning algorithms (such as Q-learning, deep deterministic policy gradient DDPG) to optimize the weights.

[0102] The training samples of the fuzzy neural network compensation model include: the input training vector includes the pressure deviation ΔP, the pressure deviation rate ΔP', the billet yield strength σ s , the extrusion ratio of the billet λ, and the real-time temperature T of the billet; the output training vector includes the compensation coefficient K c ; the training algorithm adopts the gradient descent method with momentum term, the learning rate η = 0.01~0.1, the momentum factor b = 0.8~0.95, and the error function , where m is the number of samples, represents the compensation coefficient predicted by the model for the ith sample, represents the real compensation coefficient corresponding to the ith sample.

[0103] This cold extrusion forming dynamic pressure compensation process can compensate for the first pressure head 22 or the second pressure head 23, so that the low-carbon steel has better forming effect in cold extrusion forming.

[0104] The above merely describes preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art, according to the technical solution and inventive concept of the present application, makes equivalent replacement or change within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A low-carbon steel cold extrusion forming production line, comprising a feeding conveying line (3), a hydraulic machine main body (5) composed of an upper cross beam, a movable beam and a lower cross beam, a hydraulic system, a main cylinder, an extrusion mechanism, a finished product conveying line (7) and an upper computer (2), the upper computer (2) is in communication connection with the feeding conveying line (3), the hydraulic machine main body (5), the finished product conveying line (7) and the extrusion mechanism, characterized in that, The feeding conveying line (3) and the finished product conveying line (7) are arranged on one side of the hydraulic machine body (5), the hydraulic machine body (5) is internally provided with an extrusion mechanism, and the extrusion mechanism comprises a lower die rotating mechanism (9), a pressure head switching mechanism (10) and a stress base (11); The pressure head switching mechanism (10) comprises a pushing oil cylinder (19), a backing plate (20), a sliding plate (21), a first pressure head (22) and a second pressure head (23), the output end of the pushing oil cylinder (19) is fixedly connected to one side of the sliding plate (21), the lower plate surface of the sliding plate (21) is fixedly connected to the first pressure head (22) and the second pressure head (23), the sliding plate (21) slides on the lower plate surface of the backing plate (20), and the backing plate (20) is fixedly connected below the movable beam; The lower die rotating mechanism (9) comprises a rotating mechanism and at least one lower die, the rotating mechanism drives the lower die to rotate, and the lower die corresponds to the stress base (11), the first pressure head (22) or the second pressure head (23) when being pressed. The pressure head switching mechanism (10) is used for executing a switching action when the lower die rotates to a target station.

2. A low carbon steel cold extrusion forming line according to claim 1, characterized in that, The lower die comprises a die cavity (151), a die frame and an ejection rod (152), and the die cavity (151) corresponds to the pressure head switching mechanism (10); The lower die has four lower dies, namely a first lower die (15), a second lower die (16), a third lower die (17) and a fourth lower die (18), and the first lower die (15), the second lower die (16), the third lower die (17) and the fourth lower die (18) are sequentially annular and can be pressed by the pressure head switching mechanism (10).

3. A cold extrusion forming dynamic pressure compensation process applied to the low-carbon steel cold extrusion forming production line of any one of the preceding claims 1-2, comprising the following process steps: S1, data acquisition and preprocessing; S2, generating a theoretical extrusion pressure curve; S3, monitoring the actual pressure and calculating the pressure deviation; S4, controlling the fuzzy neural network compensation model to output a compensation amount; S5, PID adaptive parameter dynamic adjustment; S6, hydraulic system dynamic response step S5; S7, compensation effect evaluation and iterative optimization.

4. The cold extrusion forming dynamic pressure compensation process according to claim 3, wherein, in S7, the compensation effect is judged based on the evaluation index, and if the requirement is not met, the model parameters or control strategy are optimized in the reverse direction to form a closed-loop evolution; The data acquisition in S1 includes collecting the physical parameters of the billet, including billet density ρ, billet hardness HV, and billet yield strength σ. s The billet's real-time temperature T, the lower die's geometric parameters (including the lower die extrusion ratio λ and the lower die angle α), and the equipment's status parameters (including the hydraulic system stiffness K) are all important parameters. s Pump flow rate Q; The S1 further comprises establishing a blank constitutive relation model Wherein, σ is the flow stress of the blank, σ0 is the strength coefficient of the blank, n is the strain hardening index of the blank, C is the strain rate sensitivity coefficient of the blank, ε is the equivalent strain of the blank, and ε´ is the equivalent strain rate of the blank.

5. A cold extrusion forming dynamic pressure compensation process according to claim 3, characterised in that, In the S2, the extrusion pressure P is calculated based on a multi-field coupling model th , and the calculation formula is , wherein K1 is a three-way pressure stress correction coefficient, σ s is the yield strength of the blank, λ is the extrusion ratio of the blank, K2 is a nonlinear index of the extrusion ratio, μ is the friction coefficient of the contact surface between the lower die and the blank, α is the die angle of the lower die, ΔT is the temperature change of the blank, and K3 is a temperature sensitivity coefficient.

6. A cold-swage forming dynamic pressure compensation process according to claim 3, characterised in that, In the S3, the actual pressure P is collected in real time by the pressure sensor embedded in the first pressure head (22) or the second pressure head (23) act , and the pressure deviation ΔP between the actual pressure and the theoretical extrusion pressure is calculated, and the calculation formula is: , wherein P th represents the theoretical extrusion pressure, and P act represents the actual pressure; the pressure deviation change rate calculation formula is: , wherein ΔP(t) represents the current time pressure deviation, ΔP(t-1) represents the previous time pressure deviation, and Δt represents the sampling time interval.

7. A cold-swage forming dynamic pressure compensation process according to claim 3, wherein, In the S4, the pressure deviation ΔP, the pressure deviation change rate ΔP', the blank yield strength σ s , the extrusion ratio λ of the blank, and the real-time temperature T of the blank are used as the input layer variables of the fuzzy neural network compensation model.

8. A cold-swage forming dynamic pressure compensation process according to claim 3, wherein, In the S6, the hydraulic system receives the PID control signal of the step S5, and adjusts the valve opening degree , and the calculation formula is , wherein, is the initial opening degree, K q is the valve gain, and ΔU is the compensation amount.

9. A cold-swage forming dynamic pressure compensation process according to claim 3, wherein, The evaluation index comprises an average absolute percentage error MAPE evaluation and a dynamic performance index evaluation, Based on the actual pressure data collected in S3 and the theoretical extrusion pressure curve in S2, the average absolute percentage error is calculated, and the calculation formula is: i is the index of the sampling point (from 1 to N); When the mean absolute percentage error MAPE exceeds the threshold value, the model parameter self-learning mechanism is triggered to update the blank constitutive model and neural network weights, wherein: P thi is the theoretical extrusion pressure value at the i-th sampling point (theoretical extrusion pressure curve from step S2); P acti is the actual extrusion pressure value of the i-th sampling point (acquired from the pressure sensor in step S3); N is the total number of sampling points participating in the calculation (i.e. the number of pressure data collected in the evaluation period). The threshold value of the average absolute percentage error MAPE is set to 3%, and when the average absolute percentage error MAPE is greater than 3%, the calculation formula of the neural network weight value of the model parameter self-learning mechanism is updated as follows:

10. A cold extrusion dynamic pressure compensation process according to claim 9, characterised in that, η represents the learning rate, wherein, denotes the current weight, α represents the learning rate, δ j denotes an error term, Indicates input layer variables, ​ Wt+1ij represents the weight of the connection between the "input layer i-th neuron" and the "hidden layer j-th neuron" at the t+1 iteration in the neural network.

Images