Intelligent press based on electric push rod driving

Abstract

The invention discloses an intelligent press based on electric push rod driving, which comprises a press base, a workbench is arranged on the press base, a pressing assembly is arranged above the press base, and the pressing assembly comprises a control cabinet body, a direct-current electric push rod serving as a press fitting execution mechanism, a battery pack, a charger and an intelligent electric press control system. The direct-current electric push rod has the capacity of outputting high-precision axial force signals and position signals in real time, the direct-current electric push rod, the battery pack and the charger are arranged in the control cabinet body, a push rod body of the direct-current electric push rod extends out of the control cabinet body and then is located over the workbench, a touch screen is arranged in front of the control cabinet body, and the direct-current electric push rod is electrically connected with the battery pack. The battery pack is electrically connected with the charger, a sound loudspeaker is arranged on the side edge of the control cabinet body, the intelligent electric press control system comprises a multi-mode process sensing module and a controller, and the controller is electrically connected with the multi-mode process sensing module, the sound loudspeaker, the direct-current electric push rod and the battery pack. The device is small in size and low in equipment cost.

Description

Technical Field

[0001] This invention relates to the technical field of presses, specifically to an intelligent press based on electric push rod drive. Background Technology

[0002] A press is a molding machine used to form industrial products by pressure. The most common type is the servo press (such as the intelligent servo press disclosed in patent application number 201510234638.6). Servo presses typically use servo motors for drive and control. Servo-driven presses have the following problems: 1. The cost of the required servo cylinders and corresponding servo systems is high, resulting in a high overall cost, at least tens of thousands of yuan; 2. The large number of components and the need for external configurations make assembly inconvenient and the structure relatively bulky; 3. The use of precision servo motors, encoders, and other components in servo cylinders leads to more potential failure points, requiring professional inspection and maintenance, which undoubtedly increases equipment maintenance costs; 4. High power consumption; 5. The use of precision servo motors, encoders, and other components in servo cylinders requires high precision and results in poor adaptability to working environments with high dust and humidity, thus requiring improvement. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent press based on electric actuator drive to solve the problems mentioned in the background art, such as high cost of the entire press, inconvenient assembly of the overall structure, large size, many failure points in the later stage, requiring professional personnel for inspection and maintenance, high power consumption, high relative precision requirements, and poor adaptability to working environments with a lot of dust and high humidity.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent press based on electric actuator drive, comprising a press base, a worktable mounted on the press base, and a pressing assembly positioned above the worktable on the press base. The pressing assembly includes a control cabinet, a DC electric actuator serving as the pressing mechanism, a battery pack, a charger, and an intelligent electric press control system. The DC electric actuator has the capability to output high-precision axial force and position signals in real time. The DC electric actuator, battery pack, and charger are housed within the control cabinet. The actuator on the DC electric actuator extends out of the control cabinet and is positioned directly above the worktable. A touchscreen is located at the front of the control cabinet. The DC electric actuator is electrically connected to the battery pack, and the battery pack is electrically connected to the charger. A speaker is located on the side of the control cabinet. The intelligent electric press control system includes: A multimodal process sensing module, integrated at the end of the DC electric push rod or on the pressure head, includes at least a triaxial force sensor and an acoustic emission sensor, wherein the triaxial force sensor is used to sense lateral force and torque; The controller, which has an embedded standard process database, a real-time process curve comparison engine, and an adaptive optimization algorithm module, is installed inside the control cabinet. The controller is electrically connected to the multimodal process sensing module, the speaker, the DC electric actuator, and the battery pack.

[0005] Preferably, the device also includes a pressure head base disposed at the push rod end of the DC electric actuator. One end of the pressure head base is provided with a first interface, and the other end of the pressure head base is provided with a second interface for installing a replaceable pressure head sleeve. The first interface is used to connect to the output flange at the push rod end of the DC electric actuator. Inside the pressure head base, there is an integrated triaxial force sensing unit compartment, which is located on the main force flow axis between the first interface and the second interface. The triaxial force sensor is installed in the triaxial force sensing unit compartment. An acoustic emission sensor mounting cavity is provided inside the side wall of the pressure head base, and the acoustic emission sensor is mounted inside the acoustic emission sensor mounting cavity. Rigid acoustic coupling is achieved through a built-in mechanical waveguide structure with the outer wall of the triaxial force sensing unit compartment or the force-bearing surface of the second interface. The output cables of the triaxial force sensor and the acoustic emission sensor are routed through the wiring groove inside the pressure head base and converge into a multi-functional aviation plug to achieve a single-wire harness connection with the external controller.

[0006] Preferably, it also includes a cloud-based process brain for storing and analyzing massive amounts of press-fitting process curves, training and distributing optimized process models; and communicating with one or more of the controllers, characterized in that the cloud-based process brain includes the following modules: The press-fit data storage module is used to receive and store historical press-fit process curve data and related production metadata from each controller; The model analysis module, based on stored process data, trains or optimizes the qualification judgment model and parameter optimization model of the press-fitting process through machine learning algorithms. The data transmission module is used to distribute the updated qualification judgment model or parameter optimization model to the relevant controllers. This invention realizes a paradigm shift in the press-fitting process from "experience-dependent" to "data-driven," enabling process knowledge to be accumulated, iterated, and passed on. It breaks through the learning capacity limit of a single device, and by aggregating data from the entire network, it can discover rare defect patterns and microscopic correlations, achieving collective intelligence where "one machine learns, the whole network benefits." It significantly reduces reliance on senior process engineers; production changeovers and parameter adjustments can be quickly completed by calling cloud-optimized process packages, greatly improving production flexibility.

[0007] Preferably, the battery pack is mounted on a Z-shaped mounting plate, which is fixed inside the control cabinet.

[0008] Preferably, to enhance the protection of the touchscreen, a screen protective cover is hinged to the touchscreen.

[0009] Preferably, the front end of the control cabinet is equipped with an equipment malfunction indicator light, an equipment normal indicator light, a power switch button, and a power-on reset button, and a cooling fan is also provided on the side of the control cabinet.

[0010] Preferably, the press base includes a press base, a press top cover, and five columns. The press base and the press top cover are arranged in parallel, and the five columns are connected between the press base and the press top cover. The control cabinet is installed on the press top cover. The push rod on the DC electric push rod passes through the press top cover. The columns are threadedly connected to the press top cover by a screw and to the press base by a screw. The press base has one or more "U"-shaped notches on both sides.

[0011] The present invention also discloses an intelligent pressing method based on the aforementioned intelligent press, which includes the following steps: S1. Online generation and comparison steps of the pressing process: During each pressing process, a multi-dimensional process curve is plotted in real time with time as the horizontal axis and axial pressure, lateral force, and acoustic emission signal as the vertical axis; the curve is compared with the pre-stored standard qualified process curve window in real time for similarity. S2. Intelligent decision-making steps based on process state machine: The pressing strategy is further dynamically adjusted according to the real-time comparison results and the preset process state machine; the process state machine includes at least the states of "normal pressing", "obstruction and hesitation", "abnormal friction" and "in place fit", and each state corresponds to different control parameters (such as speed-force switching point, pressure holding time). S3. Self-learning and press-fitting optimization steps: For press-fitting processes that are judged to be qualified but have characteristic differences in the process curve, record the differences and operating conditions (such as batch number and ambient temperature), and upload them to the cloud-based process brain; the cloud uses big data analysis to iteratively update the boundaries of the standard qualified process curve window or generate sub-process models for different operating conditions, and then send them to the controller.

[0012] Compared with the prior art, the beneficial effects of the present invention are: 1. Lower cost The overall procurement cost and the cost of the matching driver for DC electric actuators are much lower than those for servo electric cylinders and corresponding servo systems, and the price of maintenance parts is also more affordable.

[0013] 2. Simpler structure and easier installation This invention has a high degree of integration, requires no complex peripheral configuration, has a compact body size, requires less installation space for the press, and has a simpler assembly process.

[0014] 3. Low maintenance difficulty Compared to the precision servo motors, encoders, and other components of servo cylinders, DC electric actuators have a simpler mechanical structure and control logic, fewer potential points of failure, and can be inspected and maintained by non-professionals.

[0015] 4. Relatively economical energy consumption In low-load, low-frequency operating scenarios of the press, DC electric actuators have better power consumption performance and can save certain operating costs in the long run.

[0016] 5. More adaptable to different working conditions For presses operating in environments with high dust levels and high humidity, the protective design of DC electric actuators is more likely to meet the requirements and is less prone to accuracy drift or malfunction due to environmental factors.

[0017] 6. Appropriate precision For applications where high precision is not required (repeat positioning accuracy ≥ 0.05mm), DC motors can achieve this by adding potentiometers, Hall sensors, photoelectric sensors, etc., thus avoiding the precision redundancy of servo cylinders and achieving a cost-effective solution. 7. Achieve intelligent pressing and data traceability functions. Furthermore, this invention introduces a multimodal process perception module, which compares and judges multi-dimensional process curves with preset standard qualified process curves, thereby achieving accuracy and efficiency in the pressing process, greatly improving the consistency of pressing quality and first-pass yield, realizing zero-defect production, and using digital archives to meet stringent industry traceability requirements; further reducing changeover costs with greater flexibility; and combining the local perception control and calculation of the intelligent push rod with cloud big data iterative optimization to build a continuously evolving "distributed intelligent pressing system" that realizes automated data traceability function. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the structure of an intelligent press without a mounting head base based on electric push rod drive in Embodiment 1. Figure 2 This is a schematic diagram of the structure of an intelligent press based on electric push rod drive in Embodiment 1 when it is separated from the press head base; Figure 3 for Figure 1 A structural diagram without an installed output flange; Figure 4 for Figure 3 The front view; Figure 5 for Figure 3 Structural diagram without a control cabinet installed; Figure 6 This is a schematic diagram of the internal structure of the pressure head base in Embodiment 1. Figure 7 A connection diagram for the controller, multimodal process sensing module, speaker, DC electric actuator, battery pack, equipment malfunction indicator light, equipment normal indicator light, power switch button, power-on reset button, cooling fan, multimodal process sensing module, and cloud-based process brain; Figure 8 for Figure 3 A schematic diagram of the structure without the mounting base of the press.

[0019] In the diagram: 1. Press base; 2. Workbench; 3. Control cabinet; 4. DC electric actuator; 5. Battery pack; 6. Charger; 7. Touch screen; 8. Speaker; 9. Multimodal process sensing module; 10. Triaxial force sensor; 11. Acoustic emission sensor; 12. Controller; 13. Z-shaped mounting plate; 14. Screen protective cover; 15. Equipment malfunction indicator light; 16. Equipment normal indicator light; 17. Power switch button; 18. Power-on reset button; 19. Cooling fan; 101. Press base; 102. Press top cover; 103. Column; 104. Screw 1; 105. Notch; 106. Press head base; 11. First interface; 111. Second interface; 112. Internal thread; 1121. Triaxial force sensing unit compartment; 113. Acoustic emission sensor mounting cavity; 114. Wiring trough; 115. Multifunctional aviation plug; 116. Cloud-based process brain; 19. Press data storage module; 191. Model analysis module; 192. Data transmission module; 193. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0022] Example 1 Please see Figures 1-8 As shown, this embodiment discloses an intelligent press based on electric push rod drive, including a press base 1, a worktable 2 on the press base 1, and a pressing assembly located above the worktable 2. The pressing assembly includes a control cabinet 3, a DC electric push rod 4 as a pressing actuator, a battery pack 5, a charger 6, and an intelligent electric press control system. The DC electric push rod 4 has the ability to output high-precision axial force and position signals in real time. The DC electric push rod 4, battery pack 5, and charger 6 are located inside the control cabinet 3. The push rod on the DC electric push rod 4 extends out of the control cabinet 3 and is located directly above the worktable 2. A touch screen 7 is located at the front of the control cabinet 3. The DC electric push rod 4 is electrically connected to the battery pack 5, and the battery pack 5 is electrically connected to the charger 6. A speaker 8 is located on the side of the control cabinet 3. The intelligent electric press control system includes: The multimodal process sensing module 9 is integrated at the end of the DC electric push rod 4 or on the pressure head, and includes at least a triaxial force sensor 901 and an acoustic emission sensor 902. The triaxial force sensor 901 is used to sense lateral force and torque. The controller 10 has an embedded standard process database, a real-time process curve comparison engine, and an adaptive optimization algorithm module. The controller 10 is installed inside the control cabinet 3. The controller 10 is electrically connected to the multimodal process sensing module 9, the speaker 8, the DC electric actuator 4, and the battery pack 5.

[0023] Preferably, the device also includes a pressure head base 11 disposed at the push rod end of the DC electric push rod 4. One end of the pressure head base 11 is provided with a first interface 111, and the other end of the pressure head base 11 is provided with a second interface 112 for installing a replaceable pressure head sleeve. The first interface 111 is used to connect to the output flange 401 at the push rod end of the DC electric push rod 4. In this embodiment, the second interface 112 is provided with an internal thread 1121. The internal thread 1121 is used to connect with the external thread on the top of the cylindrical pressure head sleeve. Then, different types of pressure heads can be installed in the pressure head sleeve to realize the pressing process of different products. Inside the pressure head base 11, there is an integrated triaxial force sensing unit chamber 113. The triaxial force sensing unit chamber 113 is located on the main force flow axis between the first interface 111 and the second interface 112. The triaxial force sensor 901 is installed in the triaxial force sensing unit chamber 113. An acoustic emission sensor mounting cavity 114 is provided inside the side wall of the pressure head base 11, and the acoustic emission sensor 902 is installed in the acoustic emission sensor mounting cavity 114. Rigid acoustic coupling is achieved through a built-in mechanical waveguide structure with the outer wall of the triaxial force sensing unit compartment 113 or the force-bearing surface of the second interface 112. The output cables of the triaxial force sensor 901 and the acoustic emission sensor 902 pass through the wiring groove 115 inside the pressure head base 11 and converge into a multi-functional aviation plug 116, realizing a single-wire harness connection with the external controller 10. The acoustic emission sensor 902 is used to capture stress wave signals in the press-fit contact area, and the pressure head base 11 is made of a robust columnar or disc-shaped material. One end of the pressure head base 11 is fixedly connected to the output flange 401 by screws, and the other end is equipped with the pressure head. The triaxial force sensor 901 is pre-tightened and installed in the core area inside the pressure head base 11 with high precision. Its measuring axis is strictly coincident with the mechanical axis of the push rod and is located on the core path of force flow. It can directly measure all axial pressure (Fz) during the pressing process without loss, as well as the lateral force (Fx, Fy) and the axial torque (Mz) caused by workpiece misalignment and mold misalignment. This achieves signal purity and avoids the complex coupling signals introduced by structural deformation and bolt pre-tightening force changes when the sensor is installed on the push rod housing or external bracket. The pressure head base 11 ultimately constitutes an "intelligent force sensing adapter" as a standard functional module that adapts to different models of push rods and pressure heads, which facilitates product serialization. In the later stage, the triaxial force sensor 901 is installed by pre-tightening bolt group and positioning pin to ensure that its measuring center coincides with the axis of the push rod.

[0024] Preferably, it also includes a cloud-based process brain 19 for storing and analyzing massive amounts of press-fit process curves, training and distributing optimized process models; and is communicatively connected to one or more of the controllers 10. The cloud-based process brain 19 comprises the following modules: The press-fit data storage module 191 is used to receive and store historical press-fit process curve data and related production metadata from each controller 10; The model analysis module 192, based on the process data stored in the pressing data storage module 191, trains or optimizes the qualification judgment model and parameter optimization model of the pressing process through machine learning algorithms. The data sending module 193 is used to send the updated qualification judgment model or parameter optimization model to the relevant controller 10. The model analysis module 192 is connected to both the data sending module 193 and the press-fit data storage module 191. This embodiment realizes the transformation of the press-fit process from "experience-dependent" to "data-driven," enabling process knowledge to be accumulated, iterated, inherited, and traced. It breaks through the learning capacity limit of a single device, and by aggregating data from the entire network, it can discover rare defect patterns and micro-correlations, achieving collective intelligence where "one machine learns, the whole network benefits." It significantly reduces reliance on senior process engineers, and production changeovers and parameter adjustments can be quickly completed by calling the optimized process package from the cloud, greatly improving production flexibility.

[0025] Preferably, the battery pack 5 is mounted on a Z-shaped mounting plate 12, which is fixed inside the control cabinet 3.

[0026] As a preferred option, to enhance the protection of the touchscreen, a screen cover 13 is hinged to the touchscreen 7. By adding the screen cover 13, the protection of the touchscreen 7 is enhanced when the touchscreen 7 is not in use.

[0027] Preferably, the front end of the control cabinet 3 is provided with an equipment abnormality indicator light 14, an equipment normal indicator light 15, a power switch button 16, and a power-on reset button 17. The side of the control cabinet 3 is also provided with a cooling fan 18. By adding the cooling fan 18, the internal components of the control cabinet 3 can be cooled, further preventing damage to internal components due to excessive internal temperature.

[0028] Preferably, the press base 1 includes a press base 101, a press top cover 102, and five columns 103. The press base 101 and the press top cover 102 are arranged in parallel, and the five columns 103 are connected between the press base 101 and the press top cover 102. The control cabinet 3 is installed on the press top cover 102. The push rod on the DC electric push rod 4 passes through the press top cover 102. The columns 103 are threadedly connected to the press top cover 102 by screw 104, and the columns 103 are threadedly connected to the press base 101 by screw 105. The press base 101 has one or more "U"-shaped notches 106 on both sides, and the press base 101 has a control button 107 at the front end. By setting the notch 106, it is convenient to add clamps on both sides of the worktable 2 to clamp the products that need to be pressed on the worktable 2. Moreover, the clamps can be adjusted and replaced according to the products being clamped, further improving the versatility of the equipment.

[0029] This embodiment also discloses an intelligent pressing method based on the intelligent press, which includes the following steps: S1. Online generation and comparison steps of the pressing process: During each pressing process, a multi-dimensional process curve is plotted in real time with time as the horizontal axis and axial pressure, lateral force, and acoustic emission signal as the vertical axis; the curve is compared with the pre-stored standard qualified process curve window in real time for similarity. S2. Intelligent decision-making steps based on process state machine: The pressing strategy is further dynamically adjusted according to the real-time comparison results and the preset process state machine; the process state machine includes at least the states of "normal pressing", "obstruction and hesitation", "abnormal friction" and "in place fit", and each state corresponds to different control parameters (such as speed-force switching point, pressure holding time). S3. Self-learning and press-fitting optimization steps: For press-fitting processes that are judged to be qualified but have characteristic differences in the process curve, record the differences and operating conditions (such as batch number and ambient temperature), and upload them to the cloud process brain 19; The cloud uses big data analysis to iteratively update the boundary of the standard qualified process curve window or generate sub-process models for different operating conditions, and send them to the controller 10.

[0030] In this embodiment, the specific steps of the online generation and comparison step of the pressing process are as follows: S101: Real-time synchronous acquisition of multimodal time-series data of the pressing process, wherein the multimodal time-series data includes axial pressure, displacement and acoustic emission signals; S102: Based on the variation characteristics of axial pressure and displacement, the press-fitting process is divided into multiple continuous process stages online; S103: For each of the aforementioned process stages, extract a set of predefined multimodal features from the corresponding multimodal time-series data to generate a process fingerprint vector for characterizing a single press-fitting process. S104: The process fingerprint vector is compared with the pre-stored reference fingerprint model in real time, and the similarity or deviation is calculated, wherein the reference fingerprint model is established based on historical qualified pressing data. S105: Based on the similarity or deviation, output a real-time status process label to indicate the current press-fitting process status. S106: Based on the real-time status process label, dynamically adjust the control strategy of the pressing actuator; wherein, when the label is "abnormal", the adjustment of the control strategy includes at least one of the following: switching to force holding mode, performing a micro-retraction action, or safely stopping the pressing process; and the "abnormal friction" state is determined by combining the sudden energy of the acoustic emission signal in a specific frequency band with the high-frequency jitter of the axial force; the "obstruction and hesitation" state is determined by the axial force remaining at a high level while the position remains unchanged for a long time. In step S102, the pressing process is phase-divided based on predefined rules, which include at least one of the following: the moment when the axial pressure first exceeds the contact threshold, the moment when the displacement reaches the target value, and the moment when the axial pressure enters the steady-state holding range. In step S103, the extracted multimodal features include the following two types: (1) The time-domain and shape characteristics of the axial pressure and displacement curves, including average force, maximum force, force rise slope, area under the curve, and dynamic time warping distance from the reference template curve; (2) Energy and frequency domain characteristics of the acoustic emission signal, including root mean square value, absolute energy, and energy percentage in one or more predetermined frequency bands.

[0031] In step S104, the reference fingerprint model is a statistical model that defines the dynamic boundary of a multidimensional feature space formed by historical qualified data; and the specific method for real-time comparison is to calculate the Mahalanobis distance or standard deviation multiple of the process fingerprint vector relative to the dynamic boundary of the multidimensional feature space. In step S105, the real-time status process label includes three levels: "Normal," "Process Fluctuation," and "Abnormal." Furthermore, when the label is "Process Fluctuation" or "Abnormal," the method further includes step S107: based on the feature dimensions of specific deviations in the process fingerprint vector, the potential causes of failure are inferred in real time. The online process fingerprint processing system involved in the above method includes: The data acquisition and synchronization module is used to acquire synchronized multimodal time-series data. An online phase segmentation module is used to divide the process stages in real time; The process fingerprint generation module is used to extract multimodal features in stages and generate process fingerprint vectors; the real-time comparison and analysis module stores the reference fingerprint model and is used to perform comparison calculations and generate similarity or deviation results. The status decision and output module is used to output the real-time status process label based on the comparison results, which is used to directly control or adjust the operating parameters of the intelligent electric push rod actuator. The real-time comparison and analysis module is deployed in the controller 10 of the pressing equipment to perform real-time comparison with millisecond latency; the reference fingerprint model is periodically trained and updated by the cloud server and sent to the controller 10.

[0032] In this embodiment, the specific method for the intelligent decision-making steps based on process state machines includes the following steps: S201: Obtain the real-time process status label of the current pressing process, wherein the real-time process status label is generated based on the real-time comparison result of the current process fingerprint and the reference model; S202: Based on the real-time process status label, drive a preset process state machine to perform state transitions. The process state machine defines multiple discrete process states and transition conditions between states. S203: Based on the target state activated after the process state machine transition, match and call the corresponding control strategy from the preset control strategy library; S204: Based on the invoked control strategy, generate control commands and send them to the pressing actuator to dynamically adjust the pressing process.

[0033] The process states defined in the process state machine include at least the following: normal pressing state, obstruction and hesitation state, abnormal friction state, and in-place fitting state; the migration conditions include at least the real-time process state label and the real-time signal characteristics obtained from the multimodal process sensing module, and when the process state machine migrates to the obstruction and hesitation state, the control strategy invoked includes at least one of the following: (1) maintain the current output force for a preset time and monitor displacement changes; (2) perform a micro-retraction action of a preset amplitude and then try pressing again at a reduced speed; (3) trigger a "workpiece abnormality" warning. When the process state machine transitions to the abnormal friction state, the control strategy invoked includes: immediately or smoothly switching the position control mode to the force control mode, limiting the output force within a safe threshold, and triggering an audible and visual alarm; The process state machine also includes a safety protection state. The system transitions to this state when one of the following conditions is met: axial pressure exceeds the absolute safety threshold, lateral force exceeds the lateral safety threshold, or system health is below the minimum threshold. After transitioning to the safety protection state, the invoked control strategy is to execute an emergency stop and lock the actuators. The intelligent decision-making system involved in the specific method of the intelligent decision-making steps based on process state machines includes: A status input interface is used to receive the real-time process status tag and real-time signal characteristics; A process state machine engine is used to execute state transition logic based on input; Control policy mapping and executor, used to query and execute the corresponding control policy based on the activated state; The instruction output interface is used to output control instructions to the pressing actuator. The process state machine engine is a rule-based state machine or a logic judgment module implemented based on a finite automaton model. The real-time process status tag generated by the online process fingerprint processing system is input to the intelligent decision system, which then outputs control instructions to the controller 10.

[0034] In this embodiment, the specific method for the self-learning and press-fitting optimization steps includes the following steps: S301: First, receive a press-fitting process data packet from the plurality of controllers 10, the data packet containing at least a process fingerprint vector, a final quality judgment result, and production condition metadata; S302: Then, on the cloud server, based on the process fingerprint vector and quality judgment results, perform data-driven process analysis to identify process deviation patterns and potential optimization directions; S303: Based on the analysis results, update or generate an optimized process knowledge model, wherein the process knowledge model includes at least an updated qualified process fingerprint boundary model or an optimized set of press-fit control parameters; S304: The updated process knowledge model is encapsulated into a process knowledge package and sent to the controller 10 for loading and application in the subsequent pressing process.

[0035] Specifically, the data-driven process analysis in step S302 includes: S3021: Perform cluster analysis on the received qualified process fingerprint vectors to discover and define new qualified process subclusters, and update the definition of qualified fingerprint boundaries accordingly; S3022: Perform anomaly pattern mining on process fingerprint vectors marked as non-conforming, and establish association rules between specific anomaly fingerprint patterns and production condition metadata. S3023: Establish a digital twin simulation model of the pressing process based on historical qualified data. In the digital twin simulation model, with pressing quality and efficiency as the optimization objectives, use reinforcement learning or Bayesian optimization algorithm to automatically optimize at least one parameter among pressing speed, force control switching point and pressure holding time, and generate the optimized pressing control parameter set. In step S301, the process data packets received from the controller 10 follow a predetermined data selection strategy, which prioritizes uploading at least one of the following types of data: (1) Data of the entire pressing process that is determined by controller 10 to be "process fluctuation" or "abnormal"; (2) Pressing data that is judged as "qualified" but whose process fingerprint vector is located at the outer edge of the current qualified boundary confidence interval; (3) Pressing data for the first run under new materials, new molds, or new parameters; In step S303, after generating the optimized process knowledge model, a verification step is also included: the new model is used to perform offline inference on a batch of historical verification data, and the inference results are compared with known quality results. Step S304 is executed and the model is deployed only when the evaluation index of the new model is better than that of the currently deployed model.

[0036] Furthermore, the system involved in the specific methods of the self-learning and press-fitting optimization steps is deployed in a cloud server. The cloud server is communicatively connected to the controller 10, forming a distributed learning network capable of uploading process data and distributing knowledge models, which includes: The data aggregation and management module is used to receive, clean, and store process data packets from controller 10; A process analysis engine is used to perform the data-driven process analysis. The model training and optimization module is used to update or generate the process knowledge model. The knowledge distribution and version management module is used to encapsulate and distribute the process knowledge package.

[0037] Furthermore, the present invention relates to a computer-readable storage medium storing a computer program, wherein when the program is executed by a processor, it implements the specific method for the intelligent decision-making steps based on the process state machine, the specific method for the self-learning and pressing optimization steps, and the online monitoring and decision-making method for the online generation and comparison steps of the pressing process.

[0038] In this embodiment, taking the press-fitting of a precision bearing as an example, if the process curve shows that the cutting step force value is too large and accompanied by high-frequency acoustic emission during press-fitting, the system will determine it as "abnormal friction" in real time. The possible reasons are that the bearing has burrs or the journal is contaminated. At this time, the equipment will immediately trigger the "micro-retraction - cleaning reminder" strategy instead of forcibly pressing in and causing the part to be scrapped.

[0039] In this embodiment, if different models of workpieces are pressed on the same production line, after the worker scans the code, the system automatically calls the corresponding sub-process model (including curve template and control parameters) from the cloud to achieve the "zero debugging" function.

[0040] This invention replaces the existing servo press structure with a common DC electric actuator 4, resulting in the following advantages for the overall structure: 1. Lower cost The overall procurement cost and the cost of the matching driver for DC electric actuators are much lower than those for servo electric cylinders and corresponding servo systems, and the price of maintenance parts is also more affordable.

[0041] 2. Simpler structure and easier installation This invention has a high degree of integration, requires no complex peripheral configuration, has a compact body size, requires less installation space for the press, and has a simpler assembly process.

[0042] 3. Low maintenance difficulty Compared to the precision servo motors, encoders, and other components of servo cylinders, DC electric actuators have a simpler mechanical structure and control logic, fewer potential points of failure, and can be inspected and maintained by non-professionals.

[0043] 4. Relatively economical energy consumption In low-load, low-frequency operating scenarios of the press, DC electric actuators have better power consumption performance and can save certain operating costs in the long run.

[0044] 5. More adaptable to different working conditions For presses operating in environments with high dust levels and high humidity, the protective design of DC electric actuators is more likely to meet the requirements and is less prone to accuracy drift or malfunction due to environmental factors.

[0045] 6. Appropriate precision For applications where high precision is not required (repeat positioning accuracy ≥ 0.05mm), DC motors can achieve this by adding potentiometers, Hall sensors, photoelectric sensors, etc., thus avoiding the precision redundancy of servo cylinders and achieving a cost-effective solution. 7. Achieve intelligent pressing and data traceability functions. Furthermore, this invention introduces a multimodal process perception module 9, which compares and judges multi-dimensional process curves with preset standard qualified process curves, thereby achieving accuracy and efficiency in the pressing process, greatly improving the consistency of pressing quality and first-pass yield, and realizing zero-defect production. At the same time, it adopts digital archives to meet the traceability requirements of stringent industries (such as automotive and aerospace); further flexibly reduces changeover costs; and combines the local perception control and calculation of the intelligent push rod with cloud big data iterative optimization to build a continuously evolving "distributed intelligent pressing system" to realize automated data traceability function.

[0046] It should also be noted that the software program editing involved in the online generation and comparison steps of the pressing process, the intelligent decision-making steps based on the process state machine, the self-learning and pressing optimization steps, as well as the program editing of each module design and the chip rotation, are all conventional technologies in this field, and therefore will not be described in detail.

[0047] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart press based on electric push rod drive, comprising a press base (1), wherein a worktable (2) is provided on the press base (1), characterized in that: Above the press base (1) is a pressing assembly located above the workbench (2). The pressing assembly includes a control cabinet (3), a DC electric push rod (4) serving as the pressing actuator, a battery pack (5), a charger (6), and an intelligent electric press control system. The DC electric push rod (4) has the ability to output high-precision axial force and position signals in real time. The DC electric push rod (4), battery pack (5), and charger (6) are located inside the control cabinet (3). The push rod on the DC electric push rod (4) extends out of the control cabinet (3) and is located directly above the workbench (2). A touch screen (7) is located in front of the control cabinet (3). The DC electric push rod (4) is electrically connected to the battery pack (5), and the battery pack (5) is electrically connected to the charger (6). A speaker (8) is located on the side of the control cabinet (3). The intelligent electric press control system includes: The multimodal process sensing module (9) is integrated at the end of the DC electric push rod (4) or the pressure head, and includes at least a triaxial force sensor (901) and an acoustic emission sensor (902). The triaxial force sensor (901) is used to sense lateral force and torque. The controller (10) is embedded with a standard process database, a real-time process curve comparison engine and an adaptive optimization algorithm module. The controller (10) is installed in the control cabinet (3). The controller (10) is electrically connected to the multimodal process sensing module (9), the speaker (8), the DC electric push rod (4), and the battery pack (5).

2. The intelligent press based on electric actuator drive according to claim 1, characterized in that: It also includes a pressure head base (11) disposed at the push rod end of the DC electric push rod (4), one end of the pressure head base (11) is provided with a first interface (111), and the other end of the pressure head base (11) is provided with a second interface (112) for installing a replaceable pressure head sleeve; the first interface (111) is used to connect to the output flange of the push rod end of the DC electric push rod (4); Inside the pressure head base (11), there is an integrated triaxial force sensing unit compartment (113). The triaxial force sensing unit compartment (113) is located on the main force flow axis between the first interface (111) and the second interface (112). The triaxial force sensor (901) is installed in the triaxial force sensing unit compartment (113). An acoustic emission sensor mounting cavity (114) is provided inside the side wall of the pressure head base (11), and the acoustic emission sensor (902) is installed in the acoustic emission sensor mounting cavity (114); The output cables of the triaxial force sensor (901) and the acoustic emission sensor (902) pass through the wiring groove (115) inside the pressure head base (11) and converge into a multi-functional aviation plug (116) to achieve a single-wire harness connection with the external controller (10).

3. The intelligent press based on electric push rod drive according to claim 1, characterized in that: It further includes a cloud process brain (19) for storing and analyzing a large amount of press-fitting process curves, training and issuing optimized process models; and communicating with one or more of the controllers (10), wherein the cloud process brain (19) includes the following modules: A press-fitting data storage module (191) for receiving and storing historical press-fitting process curve data and related production metadata from each controller (10); A model analysis module (192) for training or optimizing a qualified determination model and a parameter optimization model of the press-fitting process based on the stored process data through machine learning algorithms; A data sending module (193) for sending the updated qualified determination model or parameter optimization model to the relevant controller (10).

4. The intelligent press based on electric push rod drive according to claim 1, characterized in that: The battery pack (5) is arranged on a Z-shaped mounting plate (12), and the Z-shaped mounting plate (12) is fixed in the control cabinet body (3).

5. The intelligent press based on electric actuator drive according to claim 1, characterized in that: A screen protective cover (13) is hinged on the touch screen (7).

6. The intelligent press based on electric push rod drive according to claim 1, characterized in that: The front end of the control cabinet body (3) is provided with an equipment abnormality indicator light (14), an equipment normal indicator light (15), a power switch button (16) and a power-on reset button (17), and a heat dissipation fan (18) is further arranged on the side of the control cabinet body (3).

7. The intelligent press based on electric actuator drive according to claim 1, characterized in that: The press base (1) includes a press base (101), a press top cover (102) and five columns (103). The press base (101) and the press top cover (102) are arranged in parallel, and the five columns (103) are connected between the press base (101) and the press top cover (102). The control cabinet body (3) is arranged on the press top cover (102). The push rod on the DC electric push rod (4) penetrates through the press top cover (102). The column (103) is threadedly connected to the press top cover (102) through a first screw (104), and the column (103) is threadedly connected to the press base (101) through a first screw (105). One or more "convex"-shaped notch grooves (106) are arranged on both sides of the press base (101).

8. An intelligent pressing method based on the intelligent press of claim 3, characterized in that, It includes the following steps: S1. Online generation and comparison step of the press-fitting process: In each press-fitting process, a multi-dimensional process curve with time as the horizontal axis and axial pressure, lateral force, and acoustic emission signal as the vertical axis is drawn in real time; and the curve is compared with a pre-stored standard qualified process curve window in real time for similarity; S2. Intelligent decision-making step based on the process state machine: Further dynamically adjust the press-fitting strategy according to the real-time comparison result and the preset process state machine; S3. Self-learning and press-fitting optimization step: For the press-fitting determined to be qualified but with characteristic differences in the process curve, record its difference points and working conditions information, and upload them to the cloud process brain (19); the cloud iteratively updates the boundary of the standard qualified process curve window or generates sub-process models for different working conditions through big data analysis, and issues them to the controller (10).

Images