A process optimization method for preventing the tail end of aluminum alloy solid core rod extrusion

Abstract

The application discloses a process optimization method for preventing the tail end of aluminum alloy solid rod extrusion from being compressed, and the method is characterized in that: a concave groove is prefabricated at the center of the tail end of a casting ingot truck to improve the end stress; a gradient temperature heating is combined with a mechanical brush roller to remove the secondary oxidation film on line; a peeling extrusion pad, a cold extrusion pad with a temperature 30-50 DEG C lower than that of the casting ingot and a resistance flow guide cavity are used to control thermodynamics and flow velocity; a closed vacuum exhaust is performed in the initial stage of extrusion; in the extrusion process, multi-source time sequence data are collected in real time to extract a characteristic vector, the vector is input into a flow field distortion prediction model to calculate a turbulent flow risk index, and in the final stage, the extrusion is performed by dynamically feeding back and adjusting the pulse pressure reduction range and the pressure maintaining time according to the index; and the equivalent length of tail end pollution is obtained based on the integral operation of the risk index at the end of the extrusion, the pressure excess thickness is dynamically optimized and generated, and the tail end is separated. The application realizes the cooperation of the physical barrier and the intelligent data closed loop, accurately blocks the metal turbulent flow channel and leaves the tail end as needed, completely eliminates the tail end defect and significantly improves the yield.

Description

Technical Field

[0001] This invention relates to the field of aluminum alloy extrusion control forming technology, and more specifically, to a process optimization method for preventing the compression tail of solid aluminum alloy bar extrusion. Background Technology

[0002] Solid aluminum alloy bars are widely used in rail transportation and machinery manufacturing, requiring extremely high internal density. In hot extrusion production of aluminum alloys, due to the significant difference in flow velocity between the inner and outer layers (the outer layer experiences greater frictional resistance and flows slower, while the inner layer flows faster), impurities such as oxides, segregates, and lubricating oil on the ingot surface easily flow back along the dead zone boundary of the extrusion cylinder at the end of extrusion, becoming trapped at the tail of the product and forming a squeeze tail defect. This is a serious internal metallurgical defect that can lead to the scrapping of the entire bar or the need to remove a large portion of the tail, severely restricting yield and production efficiency. To eliminate squeeze tails, existing technologies typically employ the following traditional processes: Ingot pre-coating and peeling extrusion: Pre-coating removes the segregated layer on the ingot surface, and a peeling extrusion pad is used to isolate surface contaminants within the extrusion cylinder during extrusion, forming a peeling ring. Retaining pressure residue and increasing tail trimming: A certain thickness of pressure residue is retained at the end of extrusion, and a longer tail is removed in subsequent processes. Linear deceleration at the end of the extrusion process: The extrusion speed is reduced manually or by setting a program at the end of the extrusion process.

[0003] However, existing technologies have revealed the following inherent limitations and technical pain points in long-term practical applications: Secondary oxide film and gas entrapment are unavoidable: Although ingots are rolled, a secondary oxide film quickly forms on the ingot surface during the subsequent gradient heating process, and existing processes lack online cleaning methods before entering the extrusion cylinder. Simultaneously, air exists between the ingot and the extrusion cylinder. During high-pressure extrusion, residual gas (containing oxygen and moisture) is highly compressed and accompanied by backflow of contaminants, easily forming difficult-to-detect "gas core shrinkage tail" or "bubble shrinkage tail." Static parameter control lacks adaptability, resulting in low yield: Traditional final deceleration and residual thickness rely on fixed parameters set manually based on experience. Due to fluctuations in temperature, microstructure uniformity, and equipment condition for each ingot furnace, fixed parameters often lead to "conservative extrusion" (residual thickness is too thick, wasting material and reducing yield) or "risky extrusion" (residual thickness is too thin, and the shrinkage tail is not completely cut off). Weak end-stage turbulence blocking ability: Conventional linear deceleration cannot effectively break the metal turbulence channels formed at the end of extrusion. Once contaminants begin to backflow, simple deceleration cannot reverse the physical flow field, and the shrinkage tail will still extend inward. Lack of thermodynamic and stress field intervention: At the end of the extrusion process, the metal at the tail end reaches extremely high temperatures due to accumulated deformation heat, making the softened metal highly susceptible to radial turbulence. Current processes do not address the thermodynamic intervention of the die and tool, or the geometric stress distribution at the ingot's end.

[0004] In summary, there is an urgent need for a process method that can achieve source purification, thermodynamic intervention and blocking, and intelligent prediction and dynamic parameter optimization based on the real-time state of the extrusion flow field, so as to completely eliminate extrusion tailing and maximize the yield. Summary of the Invention

[0005] A process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks includes the following steps: S1. Ingot preparation and end-shaped pretreatment: The aluminum alloy ingot is surface-machined to remove the surface defect layer. A chamfer is machined at the front end of the ingot after machining. A concave groove is pre-made in the center of the tail end face of the ingot. The concave groove pre-compensates for the metal loss in the center at the end of the extrusion, changes the stress distribution at the tail, and disrupts the mechanical conditions for the formation of the central tail from a geometric perspective. S2. Gradient heating and online descaling: The pre-treated ingot is heated in a gradient temperature manner, so that the temperature at the head is higher than that at the tail. The gradient temperature heating compensates for the accumulation of deformation heat at the tail. After the ingot is taken out of the furnace, before entering the extrusion cylinder, the surface of the ingot is descaled online using mechanical brush rollers to remove the secondary oxide film generated during the heating process and cut off the core "dirt source" that forms the tail shrinkage. S3. Mold and Die Configuration and Thermodynamic Preparation: Extrusion is performed using a peeling extrusion pad, the diameter of which is smaller than the inner diameter of the extrusion cylinder to form a peeling gap; the center lines of the extrusion cylinder, extrusion rod, and mold are aligned, and the inner wall of the extrusion cylinder and the surface of the ingot are kept unlubricated; an extrusion pad with an internal cooling channel is prepared, and its temperature is controlled at a cold state 30~50℃ lower than the ingot temperature; a flow-blocking and guiding cavity is set at the front end of the mold to balance the metal flow rate and reduce turbulence caused by the cross-sectional flow rate difference; S4. Establishment of closed-mouth venting and peeling extrusion: The ingot is placed into the extrusion cylinder and the extrusion pad is closed. In the early stage of the extrusion rod advance, the extrusion cylinder is evacuated and vented. This can forcibly break the formed metal turbulence channel and break the macroscopic tail into harmless microscopic traces. Then the extrusion is started, and the peeling extrusion pad is used to make the surface metal of the ingot remain on the inner wall of the extrusion cylinder to form a peeling ring. S5. Pulse Flow Field Extrusion Control Based on Data Processing: During the extrusion process, the specific pressure, extrusion rod speed, and displacement data of the extruder are collected in real time to form multi-source time series data; the multi-source time series data is filtered, denoised, and timestamped to extract the pressure fluctuation variance and velocity change rate as feature vectors; the feature vectors are input into a pre-built flow field distortion prediction model to calculate the current metal turbulence risk index; when the extrusion process enters the final stage, a pulse pressure reduction and deceleration mode is used for extrusion. The specific reduction amplitude and holding time of the pulse pressure reduction are dynamically adjusted by the turbulence risk index, that is, the cycle of instantaneous pressure reduction-holding micro-speed extrusion-pressure recovery is alternately performed, which can forcibly interrupt the formed metal turbulence channel and break the macroscopic tail into harmless microscopic traces until the extrusion ends; S6. Data-driven separation of residual material and online cleaning and tail trimming: At the end of extrusion, the equivalent length of tail contamination is obtained by integral calculation based on the turbulence risk index sequence output in the final stage, and a safety margin is superimposed to dynamically generate a residual thickness control command; according to the control command, a preset thickness of residual material is retained, and the bar stack and residual material are separated; after each ingot extrusion is completed, a cleaning pad is used to push out and clean the peeling ring remaining on the inner wall of the extrusion cylinder; the tail of the extrusion bar stack is cut off.

[0006] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S2, the pretreated ingot is heated using a gradient temperature method, so that the head temperature is higher than the tail temperature. After the ingot exits the furnace, before entering the extrusion cylinder, the surface of the ingot is descaled online using a mechanical brush roller. The gradient temperature heating makes the head temperature of the ingot 10~20℃ higher than the tail temperature. When using the mechanical brush roller for online descaling, the brush roller speed is controlled at 100~200 rpm, and it performs spiral cleaning along the surface of the ingot to avoid damaging the ingot matrix due to excessive cutting, thus ensuring the absolute purity of the ingot entering the cylinder.

[0007] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S3, the diameter of the peeling extrusion pad is smaller than the inner diameter of the extrusion cylinder to form a peeling gap; the center lines of the extrusion cylinder, extrusion rod, and mold are adjusted to coincide, as follows: The diameter of the peeling extrusion pad is 1.5~3mm smaller than the inner diameter of the extrusion cylinder to form the peeling gap; The centerline alignment error of the extrusion cylinder, extrusion rod, and mold is controlled within 0.5mm.

[0008] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S4, a vacuum evacuation and exhaust operation is performed inside the extrusion cylinder, as detailed below: The vacuuming and venting operation is carried out when the extrusion rod is advanced to 5%~10% of the ingot length. The absolute pressure of the vacuuming is controlled below 0.08MPa. After the venting is completed, the venting valve is immediately closed to increase the pressure and extrusion, thus avoiding tail bubble defects caused by high-pressure air entrapment.

[0009] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S5, the multi-source time series data is filtered, denoised, and timestamp aligned. The pressure fluctuation variance and velocity change rate are extracted as feature vectors. The specific steps are as follows: S51. Channel-by-channel filtering and noise reduction: The moving average filtering method with a set window length is used to smooth the collected original time series data of specific pressure and extrusion rod speed, respectively, to filter out high-frequency electrical noise and mechanical vibration interference. S52. Reference time axis alignment: Using the timestamp of the extrusion rod displacement data as the reference time axis, the filtered specific pressure data and velocity data are resampled to the same time resolution as the reference time axis using linear interpolation, resulting in a synchronous multi-source sequence with strictly aligned timestamps; S53. Pressure fluctuation variance extraction: Set a sliding time window on the data sequence after timestamp alignment, calculate the statistical variance of the specific pressure data in each sliding window, and use it as the pressure fluctuation variance to characterize the degree of sudden change in metal flow resistance; S54. Extraction of velocity change rate: Calculate the velocity difference between adjacent time steps after timestamp alignment, and normalize it to obtain the velocity change rate, which is used to characterize the transient response characteristics of the flow rate of the hydraulic system of the extruder. S55. Feature vector construction: The pressure fluctuation variance and the velocity change rate calculated under the same sliding time window are concatenated to construct a two-dimensional feature vector, which is used as the input of the flow field distortion prediction model.

[0010] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S5, the feature vector is input into the pre-built flow field distortion prediction model to calculate the current metal turbulence risk index, as follows: The baseline feature mean vector and baseline covariance matrix under normal extrusion conditions are retrieved from the pre-built flow field distortion prediction model. The baseline feature mean vector and baseline covariance matrix are obtained by pre-training and fitting based on historical normal extrusion multi-source time series data without tail shrinkage defects. The currently constructed feature vector is input into the model, and the Mahalanobis distance between it and the baseline feature mean vector is calculated. The calculation of the Mahalanobis distance takes into account the covariance correlation between the pressure fluctuation variance and the velocity change rate, and is used to characterize the comprehensive distortion degree of the current real-time flow field deviating from the normal flow field. The Mahalanobis distance is converted into a metal turbulence risk index ranging from 0 to 1 using a nonlinear mapping function. A baseline safety value and a limit danger value for the Mahalanobis distance are set. When the Mahalanobis distance is less than the baseline safety value, the risk index output is 0. When the Mahalanobis distance is greater than the limit danger value, the risk index output is 1. When it is between the two, it is calculated incrementally according to the nonlinear mapping function. The calculated current metal turbulence risk index is output to the logic comparator of the extruder control system in real time, serving as a direct criterion for subsequent dynamic feedback adjustment of pulse pressure reduction and holding time.

[0011] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. The specific reduction amplitude and holding time of the pulse pressure reduction in step S5 are dynamically adjusted by the turbulence risk index. The specific rules for the dynamic feedback adjustment of the turbulence risk index are as follows: A first risk threshold and a second risk threshold greater than the first risk threshold are preset; When the turbulence risk index is between the first risk threshold and the second risk threshold, the pressure reduction is controlled at 5% to 8% of the current extrusion pressure, and the pressure holding time is 1 to 2 seconds. When the turbulence risk index exceeds the second risk threshold, increase the pressure drop to 10%~15% of the current extrusion pressure and extend the pressure holding time to 3~5 seconds.

[0012] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S6, an integral operation is performed based on the turbulence risk index sequence output at the end stage. The specific calculation method is as follows: The final stage is divided into n time steps, and the turbulence risk index Ri and its duration Δ for each time step are extracted. ti Calculate the equivalent length of tail pollution. Where k is a material constant, The higher the risk, the longer the duration, and the faster the extrusion speed, the longer the calculated equivalent length of contamination. This equivalent length of tail contamination is then used to determine the corresponding extrusion speed. A safety margin of 2-5mm is added as the final allowance thickness.

[0013] The beneficial effects of this invention are as follows: It abandons the traditional, blind, fixed parameter control. By collecting multi-source data on specific pressure, velocity, and displacement in real time, it extracts characteristic vectors of pressure fluctuation variance and velocity change rate with clear physical meaning. Using a flow field distortion prediction model based on Mahalanobis distance, it accurately quantifies the invisible metal turbulence risk index (0~1) inside the extrusion cylinder, achieving a leap from "post-event verification" to "real-time prediction" of the extrusion flow field. At the end of extrusion, based on the real-time calculated turbulence risk index, it dynamically adjusts the pulsed pressure reduction mode of "instantaneous pressure reduction - pressure holding at low speed - pressure recovery." This pulsed flow field disturbance can forcibly interrupt the formed metal turbulence channel, breaking down the macroscopically continuous tail into harmless microscopic traces, greatly reducing the depth of tail defects. It also eliminates the blindness of manually estimating residual pressure. By integrating the end-stage turbulence risk index sequence, duration, and extrusion speed, the "equivalent length of tail contamination" is accurately calculated, and a safety margin is dynamically added to generate a residual thickness command. This method can "leave a tail as needed" according to the actual extrusion state of each ingot, minimizing the waste of effective material while ensuring 100% removal of contaminated metal. Compared with the traditional fixed residual process, it can significantly improve the yield of a single ingot, resulting in huge economic benefits. A concave groove is pre-fabricated in the center of the ingot tail end face to pre-compensate the metal loss in the center at the end of extrusion, changing the axial tensile stress distribution at the tail and disrupting the formation conditions of central tail shrinkage from a geometric and mechanical perspective. The innovative use of a cold extrusion pad 30-50°C lower than the ingot temperature creates a "cold shell chilling effect" when in contact with the ingot tail, increasing the deformation resistance of the tail metal and forcing the surface contaminant to flow radially outward into the peeling ring, effectively preventing the contaminant from flowing radially back inward. Attached Figure Description

[0014] Figure 1 A flowchart of a process optimization method to prevent the compression tail of solid aluminum alloy bars from being squeezed out; Detailed Implementation A process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks, the specific process is as follows: Figure 1 As shown, it includes the following steps: S1. Ingot preparation and end-shaped pretreatment: The aluminum alloy ingot is surface-machined to remove the surface defect layer. A chamfer is machined at the front end of the ingot after machining. A concave groove is pre-made in the center of the tail end face of the ingot. The concave groove pre-compensates for the metal loss in the center at the end of the extrusion, changes the stress distribution at the tail, and disrupts the mechanical conditions for the formation of the central tail from a geometric perspective. S2. Gradient heating and online descaling: The pre-treated ingot is heated in a gradient temperature manner, so that the temperature at the head is higher than that at the tail. The gradient temperature heating compensates for the accumulation of deformation heat at the tail. After the ingot is taken out of the furnace, before entering the extrusion cylinder, the surface of the ingot is descaled online using mechanical brush rollers to remove the secondary oxide film generated during the heating process and cut off the core "dirt source" that forms the tail shrinkage. S3. Mold and Die Configuration and Thermodynamic Preparation: Extrusion is performed using a peeling extrusion pad, the diameter of which is smaller than the inner diameter of the extrusion cylinder to form a peeling gap; the center lines of the extrusion cylinder, extrusion rod, and mold are aligned, and the inner wall of the extrusion cylinder and the surface of the ingot are kept unlubricated; an extrusion pad with an internal cooling channel is prepared, and its temperature is controlled at a cold state 30~50℃ lower than the ingot temperature; a flow-blocking and guiding cavity is set at the front end of the mold to balance the metal flow rate and reduce turbulence caused by the cross-sectional flow rate difference; S4. Establishment of closed-end venting and peeling extrusion: The ingot is placed into the extrusion cylinder and the extrusion pad is closed. In the initial stage of extrusion rod advancement (within the stage of advancing to 5%~10% of the ingot length), a vacuum venting operation is performed inside the extrusion cylinder, which can forcibly interrupt the formed metal turbulence channel and break the macroscopic tail into harmless microscopic traces; then extrusion is started, and the peeling extrusion pad is used to make the surface metal of the ingot remain on the inner wall of the extrusion cylinder to form a peeling ring; S5. Pulse Flow Field Extrusion Control Based on Data Processing: During the extrusion process, the specific pressure, extrusion rod speed, and displacement data of the extruder are collected in real time to form multi-source time series data; the multi-source time series data is filtered, denoised, and timestamped to extract the pressure fluctuation variance and velocity change rate as feature vectors; the feature vectors are input into a pre-built flow field distortion prediction model to calculate the current metal turbulence risk index; when the extrusion process enters the final stage, a pulse pressure reduction and deceleration mode is used for extrusion. The specific reduction amplitude and holding time of the pulse pressure reduction are dynamically adjusted by the turbulence risk index, that is, the cycle of instantaneous pressure reduction-holding micro-speed extrusion-pressure recovery is alternately performed, which can forcibly interrupt the formed metal turbulence channel and break the macroscopic tail into harmless microscopic traces until the extrusion ends; S6. Data-driven separation of residual material and online cleaning and tail trimming: At the end of extrusion, the equivalent length of tail contamination is obtained by integral calculation based on the turbulence risk index sequence output in the final stage, and a safety margin is superimposed to dynamically generate a residual thickness control command; according to the control command, a preset thickness of residual material is retained, and the bar stack and residual material are separated; after each ingot extrusion is completed, a cleaning pad is used to push out and clean the peeling ring remaining on the inner wall of the extrusion cylinder; the tail of the extrusion bar stack is cut off.

[0015] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S2, the pretreated ingot is heated using a gradient temperature method, so that the head temperature is higher than the tail temperature. After the ingot exits the furnace, before entering the extrusion cylinder, the surface of the ingot is descaled online using a mechanical brush roller. The gradient temperature heating makes the head temperature of the ingot 10~20℃ higher than the tail temperature. When using the mechanical brush roller for online descaling, the brush roller speed is controlled at 100~200 rpm, and it performs spiral cleaning along the surface of the ingot to avoid damaging the ingot matrix due to excessive cutting, thus ensuring the absolute purity of the ingot entering the cylinder.

[0016] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S3, the diameter of the peeling extrusion pad is smaller than the inner diameter of the extrusion cylinder to form a peeling gap; the center lines of the extrusion cylinder, extrusion rod, and mold are adjusted to coincide, as follows: The diameter of the peeling extrusion pad is 1.5~3mm smaller than the inner diameter of the extrusion cylinder to form the peeling gap; The centerline alignment error of the extrusion cylinder, extrusion rod, and mold is controlled within 0.5mm.

[0017] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S4, a vacuum evacuation and exhaust operation is performed inside the extrusion cylinder, as detailed below: The vacuuming and venting operation is carried out when the extrusion rod is advanced to 5%~10% of the ingot length. The absolute pressure of the vacuuming is controlled below 0.08MPa. After the venting is completed, the venting valve is immediately closed to increase the pressure and extrusion, thus avoiding tail bubble defects caused by high-pressure air entrapment.

[0018] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S5, the multi-source time series data is filtered, denoised, and timestamp aligned. The pressure fluctuation variance and velocity change rate are extracted as feature vectors. The specific steps are as follows: S51. Channel-by-channel filtering and noise reduction: The moving average filtering method with a set window length is used to smooth the collected original time series data of specific pressure and extrusion rod speed, respectively, to filter out high-frequency electrical noise and mechanical vibration interference. S52. Reference time axis alignment: Using the timestamp of the extrusion rod displacement data as the reference time axis, the filtered specific pressure data and velocity data are resampled to the same time resolution as the reference time axis using linear interpolation, resulting in a synchronous multi-source sequence with strictly aligned timestamps; S53. Pressure fluctuation variance extraction: Set a sliding time window on the data sequence after timestamp alignment, calculate the statistical variance Var(P) of the specific pressure data in each sliding window, and use it as the pressure fluctuation variance to characterize the degree of abrupt change in metal flow resistance; S54. Extraction of velocity change rate: Calculate the velocity difference of the extrusion bar between adjacent time steps after timestamp alignment, and normalize it to obtain the velocity change rate ΔV, which is used to characterize the transient response characteristics of the flow rate of the extruder hydraulic system. S55. Feature vector construction: The pressure fluctuation variance and the velocity change rate calculated under the same sliding time window are concatenated to construct a two-dimensional feature vector X=[Var(P), ΔV], which is used as the input of the flow field distortion prediction model.

[0019] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S5, the feature vector is input into the pre-built flow field distortion prediction model to calculate the current metal turbulence risk index, as follows: The baseline feature mean vector and baseline covariance matrix under normal extrusion conditions are retrieved from the pre-built flow field distortion prediction model. The baseline feature mean vector and baseline covariance matrix are obtained by pre-training and fitting based on historical normal extrusion multi-source time series data without tail shrinkage defects. The currently constructed feature vector is input into the model, and the Mahalanobis distance between it and the baseline feature mean vector is calculated. The calculation of the Mahalanobis distance takes into account the covariance correlation between the pressure fluctuation variance and the velocity change rate, and is used to characterize the comprehensive distortion degree of the current real-time flow field deviating from the normal flow field. A nonlinear mapping function is used to convert the Mahalanobis distance into a metal turbulence risk index ranging from 0 to 1; wherein, a baseline safety value for the Mahalanobis distance is set. D s With extreme danger value D d When the Mahalanobis distance D is less than the baseline safety value, the risk index output is 0; when the Mahalanobis distance D is greater than the extreme danger value, the risk index output is 1; when it is in between, it is calculated incrementally according to the nonlinear mapping function. The nonlinear mapping function is specifically a piecewise mapping model based on a smooth step function, and its specific expression is as follows: Let the currently calculated Mahalanobis distance be D, and the metal turbulence risk index be R; when D < D s When, R=0; when D s ≤ D ≤ D dWhen introducing intermediate normalized variables ,at this time The risk index is calculated using a cubic Hermite smoothing interpolation function, as shown in the formula: ; when D > D d hour, R =1, this nonlinear mapping function can ensure that the first derivative is continuous and zero at the boundary point, avoiding the control system oscillation caused by the threshold step change when calculating the risk index, and ensuring that the comprehensive distortion degree of the real-time feature vector deviating from the normal distribution is smoothly transformed into a direct criterion. The calculated current metal turbulence risk index is output to the logic comparator of the extruder control system in real time, serving as a direct criterion for subsequent dynamic feedback adjustment of pulse pressure reduction and holding time.

[0020] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. The specific reduction amplitude and holding time of the pulse pressure reduction in step S5 are dynamically adjusted by the turbulence risk index. The specific rules for the dynamic feedback adjustment of the turbulence risk index are as follows: A first risk threshold and a second risk threshold greater than the first risk threshold are preset; When the turbulence risk index is between the first risk threshold and the second risk threshold, the pressure reduction is controlled at 5% to 8% of the current extrusion pressure, and the pressure holding time is 1 to 2 seconds. When the turbulence risk index exceeds the second risk threshold, the pressure drop is increased to 10%~15% of the current extrusion pressure, and the pressure holding time is extended to 3~5 seconds. By alternating the cycle of "instantaneous pressure reduction-pressure holding micro-speed extrusion-pressure recovery" until the extrusion ends, this pulse flow field can effectively block the continuous tailing channel.

[0021] Furthermore, a process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks is proposed. In step S6, an integral operation is performed based on the turbulence risk index sequence output at the end stage. The specific calculation method is as follows: The final stage is divided into n time steps, and the turbulence risk index Ri and its duration Δ for each time step are extracted. ti Calculate the equivalent length of tail pollution. Where k is a material constant, The higher the risk, the longer the duration, and the faster the extrusion speed, the longer the calculated equivalent length of contamination. This equivalent length of tail contamination is then used to determine the corresponding extrusion speed. Add a safety margin of 2-5mm as the final residual thickness (lower limit not less than 20mm). Retain the residual thickness and separate the bar stack according to this instruction. After each ingot is extruded, use a cleaning pad to push out and clean the peeling ring remaining on the inner wall of the extrusion cylinder. Finally, cut off the tail of the bar stack (the cut length is not less than 1.5 times the diameter of the bar stack) to ensure that there are no residual defects.

[0022] Example 2 This embodiment provides a process optimization method to prevent the extrusion and compression tail of 2A12 aluminum alloy solid bar stacks (Φ120mm). The ingot size used is Φ310mm×800mm. The specific steps are as follows: S1. Ingot Preparation and End Shape Pretreatment: The Φ310mm aluminum alloy round ingot is surface-machined with a single-sided cutting depth of 2.5mm to thoroughly remove surface segregation nodules, cold shuts, and the initial oxide layer. After machining, a 45° chamfer is machined at the front end of the ingot extrusion to reduce the mold filling resistance in the early stage of extrusion. At the same time, a hemispherical concave groove is pre-made at the center of the tail end face of the ingot. The opening diameter of the concave groove is 50mm (about 16% of the ingot diameter) and the depth is 15mm (about 1.8% of the ingot length). Its purpose is that when the central metal flow leads to tensile stress at the end of extrusion, the concave groove pre-compensates for the metal loss in the central part and destroys the mechanical conditions for the formation of the central tail. S2. Gradient heating and online descaling: Gradient heating is used on the pretreated ingots, controlling the temperature of the ingot head to 450℃ and the temperature of the tail to 430℃ (the head is 20℃ higher than the tail). This temperature gradient can compensate for the accumulation of deformation heat at the tail during the extrusion process, making the overall deformation resistance of the ingot consistent and the flow more uniform. After the ingot exits the furnace, before entering the extrusion cylinder, the surface of the ingot is descaled online using mechanical brush rollers. The speed of the brush rollers is controlled at 150 rpm, and a spiral cleaning is performed along the surface of the ingot. This step forcibly removes the secondary oxide film generated during the gradient heating process, cutting off the core "dirt source" that forms the tail shrinkage. S3. Tool and Die Configuration and Thermodynamic Preparation: Extrusion is carried out using a peeling extrusion pad. The inner diameter of the extrusion cylinder is 310mm, and the diameter of the peeling extrusion pad is 307.5mm (i.e., forming a peeling gap of 2.5mm). The center lines of the extrusion cylinder, extrusion rod, and die are aligned, and the centering error is controlled within 0.4mm to prevent flow deviation. The inner wall of the extrusion cylinder and the surface of the ingot are kept in an absolutely unlubricated state. An extrusion pad with an internal cooling channel is prepared, and circulating cooling water (flow rate 10L / min) is introduced into the pad. The end face temperature is controlled at 400℃ (30℃ lower than the temperature at the tail of the ingot) to form a "cold shell effect" when it contacts the tail of the ingot, which increases the deformation resistance at the tail and forces the surface dirt to flow into the peeling ring radially outward. A funnel-shaped flow-blocking and guiding cavity is set at the front end of the die to balance the metal flow rate. S4. Establishment of closed-loop venting and peeling extrusion: Place the ingot into the extrusion cylinder and close the extrusion pad. In the initial stage of extrusion rod advancement (when it advances to 8% of the ingot length, i.e., about 64mm), perform a vacuum venting operation inside the extrusion cylinder, controlling the absolute pressure below 0.07MPa to remove residual gas and dust. Immediately after venting, close the venting valve to increase the pressure, and then start extrusion. The peeling extrusion pad leaves the surface metal of the ingot on the inner wall of the extrusion cylinder to form a complete peeling ring. S5. Pulse Flow Field Extrusion Control Based on Data Processing: During the extrusion process, the following data processing and control steps are executed: 1) Data Acquisition and Preprocessing: Real-time acquisition of extrusion press specific pressure, extrusion rod speed, and displacement data (sampling frequency 100Hz) constitutes multi-source time series data. Channel-specific filtering and noise reduction are performed: a moving average filtering method with a window length of 1 second is used to smooth the raw specific pressure and speed data separately, filtering out high-frequency electrical noise and mechanical vibration interference. Reference time axis alignment: using the timestamp of the extrusion rod displacement data as the reference axis, linear interpolation is used to resample the specific pressure and speed data to the same time resolution, resulting in a synchronized multi-source sequence. Feature vector construction: a 2-second sliding time window is set, and the statistical variance Var(P) of the specific pressure data within the window is calculated; the difference in extrusion rod speed between adjacent time steps is calculated and normalized to obtain the speed change rate ΔV, constructing a two-dimensional feature vector X=[Var(P), ΔV]. 2) Flow field distortion prediction and risk index calculation: The baseline feature mean vector and baseline covariance matrix, trained based on historical data without shrinkage defects, are retrieved from the pre-built model. The real-time feature vector is input into the model, and the Mahalanobis distance between it and the baseline mean vector is calculated. A nonlinear mapping function is used to transform the Mahalanobis distance into a metal turbulence risk index between 0 and 1. R And output it to the PLC logic comparator in real time; 3) Pulse flow field feedback control: When the extrusion process enters the final stage (when 15% of the material remains), a pulse pressure reduction and deceleration mode is adopted, with a preset first risk threshold. R 1 =0.4, second risk threshold R 2 =0.7: When 0.4 < R When the pressure drop is ≤0.7 (medium risk), the pressure reduction should be controlled at 6% of the current extrusion pressure, and the holding time should be 1.5 seconds; when R When the value is >0.7 (high risk), the pressure drop is increased to 12%, the pressure holding time is 4 seconds, and the system alternately executes the cycle of "instantaneous pressure drop - pressure holding micro-speed extrusion - pressure recovery" based on the real-time calculated R value until the extrusion ends. This pulse flow field forcibly interrupts the formed metal turbulence channel. S6. Data-driven residual separation and online tail-cutting: At the end of the extrusion process, the system divides the final stage into n time steps and extracts the turbulence risk index R corresponding to each time step. i and its duration Δ t i Extrusion speed Assuming a material constant k = 0.8, calculate the equivalent length of tail-end pollution. Assuming the system calculations at the end of this extrusion process yielded... The system automatically adds a 4mm safety margin, generating a final residual thickness control command of 32mm (20mm higher than the minimum safety limit). The main cylinder stops feeding according to the command, leaving 32mm of residual pressure. The shearing mechanism separates the bar from the residual pressure. Then, a cleaning pad is used to thoroughly push out and clean the peeling ring and dirt remaining on the inner wall of the extrusion cylinder. Finally, 180mm (i.e., 1.5 times the diameter of the bar) is cut off from the tail of the extruded Φ120mm bar to ensure that there are no residual defects.

Claims

1. A process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar stacks, characterized in that, Includes the following steps: S1. Ingot preparation and end-shaped pretreatment: The aluminum alloy ingot is surface-machined to remove the surface defect layer. A chamfer is machined at the front end of the ingot after machining. A concave groove is pre-made at the center of the tail end face of the ingot. S2. Gradient heating and online descaling: The pre-treated ingot is heated in a gradient temperature manner so that the head temperature is higher than the tail temperature; after the ingot is taken out of the furnace, before entering the extrusion cylinder, the surface of the ingot is descaled online using mechanical brush rollers to remove the secondary oxide film generated during the heating process. S3. Mold and Die Configuration and Thermodynamic Preparation: Extrusion is performed using a peeling extrusion pad, the diameter of which is smaller than the inner diameter of the extrusion cylinder to form a peeling gap; the center lines of the extrusion cylinder, extrusion rod, and mold are aligned, and the inner wall of the extrusion cylinder and the surface of the ingot are kept unlubricated; an extrusion pad with an internal cooling channel is prepared, and its temperature is controlled at a cold state 30~50℃ lower than the ingot temperature; a flow-blocking and guiding cavity is set at the front end of the mold to balance the metal flow rate; S4. Establishment of closed-end venting and peeling extrusion: The ingot is placed into the extrusion cylinder and the extrusion pad is closed. In the initial stage of the extrusion rod advance, the inside of the extrusion cylinder is evacuated and vented. Then the extrusion is started, and the peeling extrusion pad is used to make the surface metal of the ingot remain on the inner wall of the extrusion cylinder to form a peeling ring. S5. Pulse flow field extrusion control based on data processing: During the extrusion process, the specific pressure, extrusion rod speed and displacement data of the extruder are collected in real time to form multi-source time series data; the multi-source time series data is filtered and denoised and timestamped, and the pressure fluctuation variance and velocity change rate are extracted as feature vectors; the feature vectors are input into the pre-built flow field distortion prediction model to calculate the current metal turbulence risk index; When the extrusion process enters the final stage, a pulse pressure reduction and deceleration mode is used for extrusion. The specific reduction amplitude and holding time of the pulse pressure reduction are dynamically adjusted by the turbulence risk index. That is, the cycle of instantaneous pressure reduction-holding micro-extrusion-pressure recovery is alternately performed until the extrusion ends. S6. Data-driven separation of residual material and online cleaning and tail trimming: At the end of extrusion, the equivalent length of tail contamination is obtained by integral calculation based on the turbulence risk index sequence output in the final stage, and a safety margin is superimposed to dynamically generate a residual thickness control command; according to the control command, a preset thickness of residual material is retained, and the bar stack and residual material are separated; after each ingot extrusion is completed, a cleaning pad is used to push out and clean the peeling ring remaining on the inner wall of the extrusion cylinder; the tail of the extrusion bar stack is cut off.

2. The process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bars as described in claim 1, characterized in that, In step S2, the pretreated ingot is heated by a gradient temperature so that the head temperature is higher than the tail temperature. After the ingot exits the furnace, before entering the extrusion cylinder, the surface of the ingot is descaled online using a mechanical brush roller. The gradient temperature heating makes the head temperature of the ingot 10-20°C higher than the tail temperature. When the mechanical brush roller is used for online descaling, the speed of the brush roller is controlled at 100-200 rpm, and it is spirally cleaned along the surface of the ingot.

3. The process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar as described in claim 1, characterized in that, In step S3, the diameter of the peeling extrusion pad is smaller than the inner diameter of the extrusion cylinder to form a peeling gap; the center lines of the extrusion cylinder, extrusion rod, and mold are adjusted to coincide, as follows: The diameter of the peeling extrusion pad is 1.5~3mm smaller than the inner diameter of the extrusion cylinder to form the peeling gap; The centerline alignment error of the extrusion cylinder, extrusion rod, and mold is controlled within 0.5mm.

4. The process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bars as described in claim 1, characterized in that, In step S4, a vacuum evacuation and exhaust operation is performed inside the extrusion cylinder, as detailed below: The vacuuming and venting operation is carried out when the extrusion rod is advanced to 5%~10% of the ingot length. The absolute pressure of the vacuuming is controlled below 0.08MPa. After the venting is completed, the venting valve is immediately closed to increase the pressure and extrusion.

5. The process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bar as described in claim 1, characterized in that, In step S5, the multi-source time series data is filtered, denoised, and timestamp aligned. The pressure fluctuation variance and velocity change rate are extracted as feature vectors. The specific steps are as follows: S51. Channel-by-channel filtering and noise reduction: The moving average filtering method with a set window length is used to smooth the collected original time series data of specific pressure and extrusion rod speed, respectively, to filter out high-frequency electrical noise and mechanical vibration interference. S52. Reference time axis alignment: Using the timestamp of the extrusion rod displacement data as the reference time axis, the filtered specific pressure data and velocity data are resampled to the same time resolution as the reference time axis using linear interpolation, resulting in a synchronous multi-source sequence with strictly aligned timestamps; S53. Pressure fluctuation variance extraction: Set a sliding time window on the data sequence after timestamp alignment, calculate the statistical variance of the specific pressure data in each sliding window, and use it as the pressure fluctuation variance to characterize the degree of sudden change in metal flow resistance; S54. Extraction of velocity change rate: Calculate the velocity difference between adjacent time steps after timestamp alignment, and normalize it to obtain the velocity change rate, which is used to characterize the transient response characteristics of the flow rate of the hydraulic system of the extruder. S55. Feature vector construction: The pressure fluctuation variance and the velocity change rate calculated under the same sliding time window are concatenated to construct a two-dimensional feature vector, which is used as the input of the flow field distortion prediction model.

6. The process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bars as described in claim 1, characterized in that, In step S5, the feature vector is input into the pre-built flow field distortion prediction model to calculate the current metal turbulence risk index, as follows: The baseline feature mean vector and baseline covariance matrix under normal extrusion conditions are retrieved from the pre-built flow field distortion prediction model. The baseline feature mean vector and baseline covariance matrix are obtained by pre-training and fitting based on historical normal extrusion multi-source time series data without tail shrinkage defects. The currently constructed feature vector is input into the model, and the Mahalanobis distance between it and the baseline feature mean vector is calculated. The calculation of the Mahalanobis distance takes into account the covariance correlation between the pressure fluctuation variance and the velocity change rate, and is used to characterize the comprehensive distortion degree of the current real-time flow field deviating from the normal flow field. The Mahalanobis distance is converted into a metal turbulence risk index ranging from 0 to 1 using a nonlinear mapping function. A baseline safety value and a limit danger value for the Mahalanobis distance are set. When the Mahalanobis distance is less than the baseline safety value, the risk index output is 0. When the Mahalanobis distance is greater than the limit danger value, the risk index output is 1. When it is between the two, it is calculated incrementally according to the nonlinear mapping function. The calculated current metal turbulence risk index is output to the logic comparator of the extruder control system in real time, serving as a direct criterion for subsequent dynamic feedback adjustment of pulse pressure reduction and holding time.

7. The process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bars as described in claim 1, characterized in that, The specific reduction amplitude and holding time of the pulse pressure reduction in step S5 are dynamically adjusted by the turbulence risk index. The specific rules for the dynamic feedback adjustment of the turbulence risk index are as follows: A first risk threshold and a second risk threshold greater than the first risk threshold are preset; When the turbulence risk index is between the first risk threshold and the second risk threshold, the pressure reduction is controlled at 5% to 8% of the current extrusion pressure, and the pressure holding time is 1 to 2 seconds. When the turbulence risk index exceeds the second risk threshold, increase the pressure drop to 10%~15% of the current extrusion pressure and extend the pressure holding time to 3~5 seconds.

8. The process optimization method for preventing the extrusion and compression tail of solid aluminum alloy bars as described in claim 1, characterized in that, In step S6, an integral operation is performed based on the turbulence risk index sequence output at the end stage. The specific calculation method is as follows: The final stage is divided into n time steps, and the turbulence risk index R corresponding to each time step is extracted. i and its duration Δ t i Calculate the equivalent length of tail pollution. Where k is a material constant, The extrusion speed corresponds to the time step, and the equivalent length of the tail contamination is calculated. A safety margin of 2-5mm is added as the final allowance thickness.

Images