Electric vehicle outer covering part mold and intelligent manufacturing method thereof

Abstract

The invention discloses an electric vehicle outer covering part mold and an intelligent manufacturing method thereof.The electric vehicle outer covering part mold comprises a fixed base, a vertical plate is fixedly installed on one side of the top end of the fixed base, a top plate is fixedly installed at the top end of the vertical plate, and a lifting air cylinder is fixedly installed in the middle of the top plate. The quantitative air cylinder drives the sealing plate to slide along the quantitative shell, electric vehicle outer covering part raw materials in the quantitative shell are quantified and then conveyed to the position between the top die and the bottom die through the pump body, the phenomenon that the electric vehicle outer covering part materials overflow or are insufficient in injection is avoided, and the qualification rate of electric vehicle outer covering part preparation is increased. By arranging the heat dissipation mechanism, the fan body extracts low-temperature gas in the external environment through the gas inlet hose, the extracted gas is conveyed to the gas spraying cover through the gas supply hose, the gas spraying cover sprays out the low-temperature gas, heat dissipation and cooling are conducted on a formed product, and the preparation efficiency of the mold for preparing the outer covering part of the electric vehicle is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of molds, and specifically to an outer covering part mold for electric vehicles and its intelligent manufacturing method. Background Art

[0002] As one of the external components of an electric vehicle body, the outer covering part of an electric vehicle has multiple uses. First of all, it plays an aesthetic role, which can increase the fashion sense and texture of the overall appearance of the electric vehicle. Secondly, it also has the function of protecting the internal mechanical structure, which can prevent mechanical components from being affected by external collisions or adverse weather and other factors. In addition, the outer covering part of an electric vehicle also has the functions of heat dissipation and sound insulation, which can effectively reduce the temperature and noise generated by mechanical components during the driving of the electric vehicle. Generally speaking, the outer covering part of an electric vehicle is an important component with multiple functions and is the guarantee for the normal operation of the electric vehicle. When producing the outer covering part of an electric vehicle, a mold is needed to prepare it. The material for the outer covering part of the electric vehicle is injected into the cavity in the mold, and the shape and size in the mold cavity are replicated to obtain the required outer covering part of the electric vehicle.

[0003] However, the traditional outer covering part mold for electric vehicles has the following disadvantages:

[0004] During the preparation process of the traditional outer covering part mold for electric vehicles, the material for preparing the outer covering part of the electric vehicle needs to be injected into the mold. If directly injected, there may be a phenomenon of overflow of the material for the outer covering part of the electric vehicle or insufficient injection of the material for the outer covering part of the electric vehicle, reducing the qualified rate of the preparation of the outer covering part of the electric vehicle. Summary of the Invention

[0005] The purpose of the present invention is to provide an outer covering part mold for electric vehicles and its intelligent manufacturing method to solve the problem that during the preparation process of the traditional outer covering part mold for electric vehicles, the material for preparing the outer covering part of the electric vehicle needs to be injected into the mold. If directly injected, there may be a phenomenon of overflow of the material for the outer covering part of the electric vehicle or insufficient injection of the material for the outer covering part of the electric vehicle, reducing the qualified rate of the preparation of the outer covering part of the electric vehicle as mentioned in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solutions: an outer covering mold for an electric vehicle, including a fixed base, on one side of the top end of the fixed base, a vertical plate is fixedly installed, on the top end of the vertical plate, a top plate is fixedly installed, in the middle of the top plate, a lifting cylinder is fixedly installed, the movable end of the lifting cylinder is fixedly installed with a lifting platform, at the bottom end of the lifting platform, a top mold is fixedly installed, on the other side of the top end of the fixed base, two limit plates are fixedly installed, between the two limit plates, a lead screw is rotatably connected, in the middle of the lead screw, a first movable block slidably connected to the fixed base is threadedly connected, on the top end of the first movable block, a demolding mechanism is provided, on the top end of the demolding mechanism, a bottom mold is fixedly installed, the top mold and the bottom mold are arranged correspondingly, at the top end of one side of the vertical plate, a metering mechanism is fixedly installed, and at the bottom end of one side of the vertical plate, a heat dissipation mechanism is fixedly installed.

[0007] As a preferred technical solution of the present invention, the demolding mechanism includes an assembly table and a demolding table. In the middle of the top end of the assembly table, a shock absorber is fixedly installed, the top end of the shock absorber is fixedly connected to the middle of the bottom end of the demolding table. On both sides of the top end of the assembly table, fixed shells are fixedly installed. On the top ends of the two fixed shells, shock absorption springs are fixedly installed. On the top ends of the two shock absorption springs, push rods are fixedly installed. The top ends of the two push rods are respectively fixedly connected to both sides of the bottom end of the demolding table. The shock absorber buffers the extrusion force of the bottom mold on the demolding table, and the push rod extrudes the shock absorption spring. The shock absorption spring has elasticity, and the shock absorption spring undergoes elastic deformation to buffer the extrusion force. The shock absorber buffers the deformation force generated by the shock absorption spring.

[0008] As a preferred technical solution of the present invention, the bottom end of the assembly table is fixedly connected to the top end of the first movable block, and the top end of the demolding table is fixedly connected to the bottom end of the bottom mold. The demolding mechanism is installed on the first movable block through the assembly table, and the demolding mechanism is installed on the bottom mold through the demolding table.

[0009] As a preferred technical solution of the present invention, the heat dissipation mechanism includes an angle seat and an angle plate. The inside of the angle seat is rotatably connected to one end of the angle plate. At the bottom end of the angle plate, an air jet cover is fixedly communicated. On the top end of the angle plate, a movable frame is fixedly installed. In the middle of the movable frame, a second movable block is slidably connected. On one side of the second movable block, a push cylinder is fixedly installed. The fixed end of the push cylinder and one side of the angle seat are both fixedly connected to the side of the vertical plate facing each other. The push cylinder performs telescopic movement. The push cylinder pushes the second movable block from one side. The second movable block slides along the movable frame, and the angle plate deflects at an angle relative to the angle seat due to the frictional force of the sliding of the second movable block, so as to adjust the heat dissipation angle of the air jet cover.

[0010] As a preferred technical solution of the present invention, a blower body is fixedly installed at the bottom end of one side of the vertical plate. An intake hose is fixedly communicated with the air inlet of the blower body, and an air supply hose extending into the inside of the air jet cover is fixedly communicated with the air outlet of the blower body. After the blower body is powered on and starts, the blower body extracts low-temperature gas from the external environment through the intake hose, and the extracted gas is transported to the inside of the air jet cover through the air supply hose. The air jet cover sprays the gas to cool and shape the prepared outer covering parts of the electric vehicle.

[0011] As a preferred technical solution of the present invention, the metering mechanism includes a metering shell and a metering frame. The top end of the metering shell is fixedly connected to the bottom end of the metering frame. A heating plate is fixedly installed inside the metering shell. A sealing plate is slidably connected to the top end inner wall of the metering shell. A vertical rod extending to the outside is fixedly installed at the top end of the sealing plate. A metering cylinder is fixedly installed in the middle of the metering frame. The movable end of the metering cylinder is fixedly connected to the side facing the vertical rod. A plurality of capacity lines are provided on the surface of the metering shell. An extraction head is fixedly communicated with the bottom end of the metering shell. One side of the metering frame is fixedly connected to the side facing the vertical plate. The metering cylinder performs telescopic movement, and the metering cylinder pushes the sealing plate to slide along the metering shell, so that the metering shell meters the raw materials of the outer covering parts of the electric vehicle.

[0012] As a preferred technical solution of the present invention, a pump body is fixedly installed on the surface of the vertical plate. An extraction hose extending into the inside of the metering shell is fixedly communicated with the liquid inlet of the pump body, and a delivery hose extending into the inside of the top mold is fixedly communicated with the liquid outlet of the pump body. After the pump body is powered on and starts, the pump body extracts the raw materials of the outer covering parts of the electric vehicle that have been metered by the metering mechanism through the extraction hose, and the extracted raw materials of the outer covering parts of the electric vehicle are transported to the space between the top mold and the bottom mold through the delivery hose.

[0013] As a preferred technical solution of the present invention, sliding frames slidably connected to the top plate are fixedly installed on both sides of the top end of the lifting table. Connecting springs are fixedly installed between the two sliding frames and the top plate. A stepping motor for driving the lead screw to rotate is fixedly installed on the surface of one of the limiting plates. During the process of the lifting cylinder pushing the lifting table, the lifting table drives the sliding frame to slide relative to the top plate, improving the stability of the lifting of the top mold. After the stepping motor is powered on and starts, the stepping motor drives the lead screw to rotate. The thread on the surface of the lead screw matches the thread on the inner wall of the first movable block. The first movable block is limited by the fixed base with a shape and size matching it, so the first movable block slides along the lead screw to adjust the position of the bottom mold.

[0014] The manufacturing method of the outer covering part mold of the electric vehicle of the present invention includes the following steps:

[0015] Step 1, raw material metering: The metering mechanism meters the materials for preparing the outer covering parts of the electric vehicle;

[0016] Step 2: Die Alignment: The first movable block slides along the lead screw so that the bottom die and the top die are longitudinally aligned;

[0017] Step 3: Die Clamping: The lifting cylinder pushes the lifting platform so that the top die is clamped inside the bottom die;

[0018] Step 4: Injection: After the pump body extracts a fixed amount of raw material, it is injected between the top die and the bottom die;

[0019] Step 5: Shaping and Cooling: The cooling mechanism extracts low-temperature gas through the fan body and blows the low-temperature gas through the air jet cover to cool and shape the product between the dies;

[0020] Step 6: Demolding and Packaging: The demolding mechanism pushes the bottom die from the bottom, strips the product from between the bottom die and the top die, and then packages it.

[0021] The specific process of Step 3 includes the following steps:

[0022] Step 1: Sensor Installation

[0023] Use a high-precision laser displacement sensor as the position sensor and install position sensors at the four corners of the top die and the bottom die to ensure that the position changes of the dies can be accurately monitored. Necessary deviation adjustments are made by comparing the sensor readings with the actual positions;

[0024] x top =[x top1 ,x top2 ,x top3 ,x top4

[0025] x bottom =[x bottom1 ,x bottom2 ,x bottom3 ,x bottom4

[0026] where x top refers to the position installed on the top die, x top1 is the first corner of the fixed die position, x top2 is the second corner of the fixed die position, x top3 is the third corner of the fixed die position, x top4 is the fourth corner of the fixed die position, x bottom refers to the position installed on the bottom film, x bottom1 is the first corner of the bottom film position, x bottom2 is the second corner of the bottom film position, x bottom3 is the third corner of the bottom film position, x bottom4 is the fourth corner of the bottom film position;

[0027] Average value of the top die position: ​​

[0028]

[0029] Mean value of the bottom film position:

[0030]

[0031] Position error:

[0032]

[0033] Wherein, is the mean value of the top mold position, x topi is the i-th corner of the top mold, is the mean value of the bottom film position, x bottomi is the i-th corner of the bottom mold, E p is the position error value;

[0034] Use a high-precision strain gauge force sensor as the force sensor, which is fixed at the top and bottom of the lifting cylinder to ensure that the sensor can accurately sense the force applied during the lifting of the cylinder. Through range calibration, apply a known force, record the sensor reading, and correct the range coefficient;

[0035] Mean value of force:

[0036]

[0037] Force error:

[0038] E f = |F top - F bottom |

[0039] Wherein, is the average force at the top and bottom, F top is the force at the top, F bottom is the force at the bottom, E f is the force error;

[0040] Use a high-precision linear encoder as the speed sensor, which is installed in the middle of the lifting cylinder to monitor the lifting speed in real time, compare the sensor reading with the actual speed, and make necessary deviation adjustments;

[0041] Speed error:

[0042] E v = |v set - v|

[0043] Where E v is the speed error, v set is the set speed, and v is the real-time speed;

[0044] Step 2: Data preprocessing

[0045] Median filtering is selected to ensure the smoothness and accuracy of the data,

[0046]

[0047] where median represents the median filtering function, x(t) represents the original data at time t, x(t - N + 1) is the data point at the earliest time within the sliding window, and N is the size of the sliding window, is the denoised data at time t;

[0048] Step 3: Model usage and control prediction

[0049] The present invention predicts and optimizes the model by using the reinforcement learning algorithm and the fuzzy control algorithm to ensure the fast and stable buckling of the top die and the bottom die. The specific operations are as follows:

[0050] First, in the simulation environment, DQN is used for policy training, and initial optimization is carried out in combination with the fuzzy control rule base. Then, the model parameters are regularly evaluated and adjusted to ensure the effectiveness of the control strategy. After training, the trained strategy is applied to the actual system and fine-tuned in combination with fuzzy control;

[0051] Reinforcement learning (DQN) part:

[0052] The parameters of the state space s include position, force, and velocity, and the specific definitions are as follows:

[0053] s = [x top ,x bottom ,F top ,F bottom ,v]

[0054] The action space a includes the control signals of the lifting cylinder:

[0055] a = [rise, fall, stop]

[0056] Rise: Increase the height of the lifting cylinder;

[0057] Fall: Decrease the height of the lifting cylinder;

[0058] Stop: Maintain the current height;

[0059] The reward function is set according to the speed, stability, and mechanical wear of the die buckling:

[0060] Reward function r t :

[0061] r t = -α·E p -β·E f -γ·E v

[0062] E p is the position error, and α, β, γ are weight coefficients used to adjust the influence of each error on the total reward, E f is the force error, E v is the speed error;

[0063] Deep Q - Network (DQN) training:

[0064] Initialization: Initialize the weights and biases of the Q - network; Initialize the experience replay buffer;

[0065] Training process: Step 1. Randomly select an initial state s0 in the simulation environment, Step 2. Select the action a at the current time t , Step 3. Execute the action and observe the reward r at the current time t and the next - time state S t+1 , Step 4. Store the experience (S t , a t , r t , S t+1 ) in the experience replay buffer, Step 5. Randomly sample a mini - batch of samples from the experience replay buffer, Step 6. Calculate the target Q - value:

[0066]

[0067] Step 7. Update the weights and biases of the Q - network using gradient descent, Step 8. Repeat Steps 2 - 7 until the maximum number of training epochs is reached;

[0068] where y is the target Q - value, γ is the discount factor, represents the action that maximizes the Q - value at the next moment, Q(s t+1 , a t+1 ) represents the expected reward that can be obtained by executing the action at the next - time state, the Q - value is used to represent the expected cumulative reward that can be obtained by taking a certain action in a given state, s t+1 is the state at the next time t + 1, a t+1 is the action at the next time t + 1;

[0069] Fuzzy control part:

[0070] Construct a fuzzy control rule base: Based on inputs such as position error and force error, output appropriate control signals; If the position error is large and the force error is small, output a "large adjustment" control signal; If the position error is small and the force error is large, output a "fine adjustment" control signal;

[0071] Fuzzy logic inference:

[0072] Input variable: Position error Ep Sum force error E f , the output variable is the control signal of the adjustment amplitude, several fuzzy sets are defined for the input variable and the output variable respectively, each fuzzy set corresponds to a membership function, and the fuzzy rules are set as follows:

[0073] If E p is large and E f is small, then large adjustment amplitude, if E p is small and E f is large, then small adjustment amplitude;

[0074] Then, according to the membership degree of the input variable, use fuzzy logic reasoning to obtain the membership degree of the output variable, and then through the defuzzification method (such as the centroid method), convert the fuzzy set into a control signal for amplitude adjustment;

[0075] Step 9: Actuator control method

[0076] Control signal generation, generate a comprehensive control signal by combining reinforcement learning and fuzzy control algorithms to ensure the optimal movement path and force of the lifting cylinder; Comprehensive control signal:

[0077] at=λ·a DQN +(1-λ)·a Fuzzy

[0078] a t is the comprehensive control signal, λ is the weight coefficient, adjusting the proportion of the two control signals, a DQN is the control signal generated by reinforcement learning, a Fuzzy is the control signal generated by fuzzy control;

[0079] Control signal transmission, transmit the generated control signal to the actuator of the lifting cylinder to achieve real-time adjustment:

[0080] Electric lifting cylinder: Adjust the height according to the input voltage signal;

[0081] Hydraulic lifting cylinder: Adjust the height according to the input hydraulic signal;

[0082] Control signal application: Adjust the movement path and force of the lifting cylinder according to the comprehensive control signal;

[0083] V=k·a t

[0084] V is the voltage signal, k is the proportionality coefficient, converting the control signal into a voltage signal;

[0085] P=k2·a t

[0086] P is a hydraulic signal, and k2 is a proportionality coefficient that converts the control signal into a hydraulic signal;

[0087] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0088] 1. By setting a metering mechanism and a pump body, the metering cylinder drives the sealing plate to slide along the metering housing, meters the raw materials for the outer coverings of electric vehicles within the metering housing, and then transports them to between the top mold and the bottom mold through the pump body, avoiding the phenomena of material overflow or insufficient injection of materials for the outer coverings of electric vehicles, and improving the qualification rate of the preparation of the outer coverings of electric vehicles;

[0089] 2. By setting a heat dissipation mechanism, the fan body extracts low-temperature gas from the external environment through the intake hose, and the extracted gas is transported to the air jet cover through the air supply hose. The air jet cover sprays out the low-temperature gas to dissipate heat and cool down the formed product, improving the preparation efficiency of the mold for preparing the outer coverings of electric vehicles.

[0090] 3. Through an intelligent control algorithm for the clamping of a mold for the outer coverings of electric vehicles, this algorithm combines reinforcement learning and fuzzy control technologies, and monitors the mold position, force, and speed data in real time through sensors to generate and dynamically adjust the control signal of the lifting cylinder. Specifically, the DQN algorithm learns the optimal control strategy by interacting with the environment, while fuzzy control is based on expert experience and rules to quickly respond to system changes. The comprehensive control signal ensures the rapid and stable clamping of the top mold and the bottom mold. Through this intelligent control algorithm, the preparation process of the mold for the outer coverings of electric vehicles is more efficient and stable, and significantly reduces the operation complexity and the need for human intervention. Description of the Drawings

[0091] Figure 1 is a side view of the present invention;

[0092] Figure 2 is a side view of the heat dissipation mechanism of the present invention;

[0093] Figure 3 is a cross-sectional view of the demolding mechanism of the present invention;

[0094] Figure 4 is a cross-sectional view of the metering mechanism of the present invention;

[0095] Figure 5 is a side view of the metering mechanism of the present invention;

[0096] Figure 6 is a flowchart of the present invention;

[0097] Figure 7 is a side view of the demolding mechanism of the present invention;

[0098] Figure 8It is a flowchart of an intelligent control algorithm for the buckling of the outer covering die of an electric vehicle.

[0099] In the figure: 1. Fixed base; 2. Limit plate; 3. First movable block; 4. Lead screw; 5. Stepper motor; 6. Demolding mechanism; 61. Assembly table; 62. Shock absorber; 63. Fixed shell; 64. Push rod; 65. Demolding table; 66. Shock spring; 7. Bottom die; 8. Heat dissipation mechanism; 81. Fan body; 82. Intake hose; 83. Angle seat; 84. Angle plate; 85. Jet hood; 86. Movable frame; 87. Push cylinder; 88. Second movable block; 89. Air supply hose; 9. Quantitative mechanism; 91. Quantitative shell; 92. Extraction head; 93. Heating plate; 94. Sealing plate; 95. Vertical rod; 96. Quantitative frame; 97. Quantitative cylinder; 98. Capacity line; 10. Extraction hose; 11. Vertical plate; 12. Top plate; 13. Sliding frame; 14. Lifting cylinder; 15. Connecting spring; 16. Lifting table; 17. Top die; 18. Delivery hose; 19. Pump body. Specific implementation mode

[0100] Next, in combination with the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0101] Please refer to Figure 1-7 , the present invention provides an outer covering die for an electric vehicle, including a fixed base 1. One side of the top of the fixed base 1 is fixedly installed with a vertical plate 11, the top of the vertical plate 11 is fixedly installed with a top plate 12, the middle of the top plate 12 is fixedly installed with a lifting cylinder 14, the movable end of the lifting cylinder 14 is fixedly installed with a lifting table 16, the bottom of the lifting table 16 is fixedly installed with a top die 17, the other side of the top of the fixed base 1 is fixedly installed with two limit plates 2, a lead screw 4 is rotatably connected between the two limit plates 2, the middle of the lead screw 4 is threadedly connected with a first movable block 3 slidably connected to the fixed base 1, a demolding mechanism 6 is arranged at the top of the first movable block 3, the top of the demolding mechanism 6 is fixedly installed with a bottom die 7, the top die 17 and the bottom die 7 are arranged correspondingly, a quantitative mechanism 9 is fixedly installed at the top of one side of the vertical plate 11, and a heat dissipation mechanism 8 is fixedly installed at the bottom of one side of the vertical plate 11.

[0102] The demolding mechanism 6 includes an assembly table 61 and a demolding table 65. A shock absorber 62 is fixedly installed in the middle of the top end of the assembly table 61. The top end of the shock absorber 62 is fixedly connected to the middle of the bottom end of the demolding table 65. Fixed shells 63 are fixedly installed on both sides of the top end of the assembly table 61. Shock-absorbing springs 66 are fixedly installed on the top ends of the two fixed shells 63. Push rods 64 are fixedly installed on the top ends of the two shock-absorbing springs 66. The top ends of the two push rods 64 are respectively fixedly connected to both sides of the bottom end of the demolding table 65. The shock absorber 62 buffers the extrusion force of the bottom mold 7 on the demolding table 65, and the push rod 64 presses the shock-absorbing spring 66. The shock-absorbing spring 66 has elasticity. The shock-absorbing spring 66 undergoes elastic deformation to buffer the extrusion force, and the shock absorber 62 buffers the deformation force generated by the shock-absorbing spring 66.

[0103] The bottom end of the assembly table 61 is fixedly connected to the top end of the first movable block 3. The top end of the demolding table 65 is fixedly connected to the bottom end of the bottom mold 7. The demolding mechanism 6 is installed on the first movable block 3 through the assembly table 61, and the demolding mechanism 6 is installed on the bottom mold 7 through the demolding table 65.

[0104] The heat dissipation mechanism 8 includes an angle seat 83 and an angle plate 84. One end of the angle plate 84 is rotatably connected to the inside of the angle seat 83. An air jet cover 85 is fixedly communicated with the bottom end of the angle plate 84. A movable frame 86 is fixedly installed at the top end of the angle plate 84. A second movable block 88 is slidably connected to the middle of the movable frame 86. A push cylinder 87 is fixedly installed on one side of the second movable block 88. The fixed end of the push cylinder 87 and one side of the angle seat 83 are both fixedly connected to the side of the vertical plate 11 facing each other. The push cylinder 87 performs telescopic movement. The push cylinder 87 pushes the second movable block 88 from one side. The second movable block 88 slides along the movable frame 86, and the angle plate 84 deflects at an angle relative to the angle seat 83 due to the frictional force of the sliding of the second movable block 88, so as to adjust the heat dissipation angle of the air jet cover 85.

[0105] A fan body 81 is fixedly installed at the bottom end of one side of the vertical plate 11. An air inlet hose 82 is fixedly communicated with the air inlet of the fan body 81. An air supply hose 89 extending into the air jet cover 85 is fixedly communicated with the air outlet of the fan body 81. After the fan body 81 is powered on and starts, the fan body 81 extracts low-temperature gas from the external environment through the air inlet hose 82, and the extracted gas is transported to the air jet cover 85 through the air supply hose 89. The air jet cover 85 sprays the gas to cool and shape the prepared electric vehicle exterior cover product.

[0106] The metering mechanism 9 includes a metering housing 91 and a metering frame 96. The top end of the metering housing 91 is fixedly connected to the bottom end of the metering frame 96. A heating plate 93 is fixedly installed inside the metering housing 91. The top end of the inner wall of the metering housing 91 is slidably connected to a sealing plate 94. The top end of the sealing plate 94 is fixedly installed with a vertical rod 95 extending to the outside. A metering cylinder 97 is fixedly installed in the middle of the metering frame 96. The movable end of the metering cylinder 97 is fixedly connected to the side facing the vertical rod 95. A plurality of capacity lines 98 are provided on the surface of the metering housing 91. The bottom end of the metering housing 91 is fixedly communicated with a suction head 92. One side of the metering frame 96 is fixedly connected to the side of the vertical plate 11 facing it. The metering cylinder 97 performs a telescopic movement, and the metering cylinder 97 pushes the sealing plate 94 to slide along the metering housing 91, so as to quantitatively meter the raw materials for the outer coverings of electric vehicles by the metering housing 91.

[0107] A pump body 19 is fixedly installed on the surface of the vertical plate 11. The liquid inlet of the pump body 19 is fixedly communicated with a suction hose 10 extending into the metering housing 91. The liquid outlet of the pump body 19 is fixedly communicated with a delivery hose 18 extending into the top mold 17. After the pump body 19 is powered on and starts, the pump body 19 sucks the raw materials for the outer coverings of electric vehicles that have been quantitatively metered by the metering mechanism 9 through the suction hose 10, and the sucked raw materials for the outer coverings of electric vehicles are transported to the space between the top mold 17 and the bottom mold 7 through the delivery hose 18.

[0108] Sliding frames 13 slidably connected to the top plate 12 are fixedly installed on both sides of the top end of the lifting table 16. Connecting springs 15 are fixedly installed between the two sliding frames 13 and the top plate 12. A stepping motor 5 for driving the lead screw 4 to rotate is fixedly installed on the surface of one of the limit plates 2. During the process of the lifting cylinder 14 pushing the lifting table 16, the lifting table 16 drives the sliding frame 13 to slide relative to the top plate 12, improving the stability of the lifting of the top mold 17. After the stepping motor 5 is powered on and starts, the stepping motor 5 drives the lead screw 4 to rotate. The thread on the surface of the lead screw 4 matches the thread on the inner wall of the first movable block 3. The first movable block 3 is limited by the fixed base 1 with a shape and size matching it, so the first movable block 3 slides along the lead screw 4 to adjust the position of the bottom mold 7.

[0109] The intelligent manufacturing method for the outer covering mold of an electric vehicle of the present invention includes the following steps:

[0110] Step 1, raw material metering: The metering mechanism 9 meters the materials for preparing the outer coverings of electric vehicles;

[0111] Step 2, mold alignment: The first movable block 3 slides along the lead screw 4 to longitudinally align the bottom mold 7 with the top mold 17;

[0112] Step 3, mold clamping: The lifting cylinder 14 pushes the lifting table 16 to clamp the top mold 17 in the bottom mold 7;

[0113] Step 4, Material Injection: After the pump body 19 extracts a certain amount of raw material, it injects the raw material between the top mold 17 and the bottom mold 7;

[0114] Step 5, Shaping and Heat Dissipation: The heat dissipation mechanism 8 extracts low-temperature gas through the fan body 81 and blows out the low-temperature gas through the air jet cover 85 to dissipate heat and shape the product between the molds;

[0115] Step 6, Demolding and Packaging: The demolding mechanism 6 pushes the bottom mold 7 from the bottom, peels the product from between the bottom mold 7 and the top mold 17, and then packages it.

[0116] In the present invention, the metering mechanism 9 meters the material for preparing the outer covering parts of the electric vehicle; after the stepping motor 5 is powered on, it starts, and the stepping motor 5 drives the lead screw 4 to rotate. The thread on the surface of the lead screw 4 matches the thread on the inner wall of the first movable block 3. The first movable block 3 is limited by the fixed base 1 that matches its shape and size. Therefore, the first movable block 3 slides along the lead screw 4 to adjust the position of the bottom mold 7. The first movable block 3 slides along the lead screw 4, making the bottom mold 7 longitudinally aligned with the top mold 17; the lifting cylinder 14 pushes the lifting platform 16, causing the top mold 17 to be buckled inside the bottom mold 7. During the process of the lifting cylinder 14 pushing the lifting platform 16, the lifting platform 16 drives the sliding frame 13 to slide relative to the top plate 12, improving the stability of the lifting of the top mold 17. The metering cylinder 97 performs a telescopic movement, and the metering cylinder 97 pushes the sealing plate 94 to slide along the metering shell 91, thereby enabling the metering shell 91 to meter the raw material for the outer covering parts of the electric vehicle. After the pump body 19 is powered on, it starts, and the pump body 19 extracts the raw material for the outer covering parts of the electric vehicle that has been metered by the metering mechanism 9 through the extraction hose 10. The extracted raw material for the outer covering parts of the electric vehicle is transported to between the top mold 17 and the bottom mold 7 through the delivery hose 18. The heat dissipation mechanism 8 extracts low-temperature gas through the fan body 81 and blows out the low-temperature gas through the air jet cover 85 to dissipate heat and shape the product between the molds; the demolding mechanism 6 pushes the bottom mold 7 from the bottom, peels the product from between the bottom mold 7 and the top mold 17, and then packages it.

[0117] In the present invention, Step 3 specifically includes the following steps:

[0118] Step 1: Sensor Installation

[0119] Use a high-precision laser displacement sensor as the position sensor, and install position sensors at the four corners of the top mold and the bottom mold to ensure that the position changes of the molds can be accurately monitored. By comparing the sensor readings with the actual positions, necessary deviation adjustments are made;

[0120] x top =[x top1 , x top2 , x top3 , x top4

[0121] x​bottom = [x bottom1 , x bottom2 , x bottom3 , x bottom4

[0122] where x top refers to the installation position on the top mold, x top1 is the first corner of the fixed mold position, x top2 is the second corner of the fixed mold position, x top3 is the third corner of the fixed mold position, x top4 is the fourth corner of the fixed mold position, x bottom refers to the installation position on the bottom mold, x bottom1 is the first corner of the bottom mold position, x bottom2 is the second corner of the bottom mold position, x bottom3 is the third corner of the bottom mold position, x bottom4 is the fourth corner of the bottom mold position;

[0123] Mean value of the top mold position:

[0124]

[0125] Mean value of the bottom mold position:

[0126]

[0127] Position error:

[0128]

[0129] where is the mean value of the top mold position, x topi is the i-th corner of the top mold, is the mean value of the bottom mold position, x bottomi is the i-th corner of the bottom mold, E p is the position error value;

[0130] Use a high-precision strain gauge force sensor as the force sensor, fixed at the top and bottom of the lifting cylinder, ensuring that the sensor can accurately sense the force applied during the lifting of the cylinder. Through range calibration, apply a known force, record the sensor readings, and correct the range coefficient;

[0131] Mean value of force:

[0132]

[0133] Force error:

[0134] E f = |F top - F bottom |

[0135] ​Among them, is the average value of the forces at the top and bottom, F top is the force at the top, F bottom is the force at the bottom, E f is the error of the force;

[0136] A high-precision linear encoder is used as a speed sensor and installed in the middle of the lifting cylinder to monitor the lifting speed in real time, compare the sensor readings with the actual speed, and make necessary deviation adjustments;

[0137] Speed error:

[0138] E v = |v set - v|

[0139] Among them, E v is the speed error, v set is the set speed, and v is the real-time speed;

[0140] Step 2: Data preprocessing

[0141] Median filtering is selected to ensure the smoothness and accuracy of the data,

[0142]

[0143] Among them, median represents the median filtering function, x(t) represents the original data at time t, x(t - N + 1) is the data point at the earliest time within the sliding window, and N is the size of the sliding window, is the denoised data at time t;

[0144] Step 3: Model use and control prediction

[0145] The present invention predicts and optimizes the model by using the reinforcement learning algorithm and the fuzzy control algorithm to ensure the fast and stable buckling of the top mold and the bottom mold. The specific operations are as follows:

[0146] First, in the simulation environment, DQN is used for policy training, and initial optimization is carried out in combination with the fuzzy control rule base. Then, the model parameters are regularly evaluated and adjusted to ensure the effectiveness of the control strategy. After training, the trained policy is applied to the actual system and fine-tuned in combination with fuzzy control;

[0147] Reinforcement learning (DQN) part:

[0148] The parameters of the state space s include position, force, and speed, and the specific definitions are as follows:

[0149] s = [x top , x bottom , F top , F bottom , v]

[0150] The action space a includes the control signal of the lifting cylinder:

[0151] a = [rise, fall, stop]

[0152] Rise: Increase the height of the lifting cylinder;

[0153] Fall: Decrease the height of the lifting cylinder;

[0154] Stop: Maintain the current height;

[0155] The reward function is set according to the mold closing speed, stability and mechanical wear:

[0156] Reward function r t :

[0157] r t = -α·E p -β·E f -γ·E v

[0158] E p is the position error, α, β, γ are weight coefficients used to adjust the influence of each error on the total reward, E f is the force error, E v is the speed error;

[0159] Deep Q-Network (DQN) training:

[0160] Initialization: Initialize the weights and biases of the Q-network; Initialize the experience replay buffer;

[0161] Training process: Step 1. Randomly select an initial state s0 in the simulation environment, Step 2. Select the action a at the current time t , Step 3. Execute the action and observe the reward r at the current time t and the next time state S t+1 , Step 4. Store the experience (S t , a t , r t , S t+1 ) in the experience replay buffer, Step 5. Randomly sample a small batch of samples from the experience replay buffer, Step 6. Calculate the target Q value:

[0162]

[0163] Step 7. Update the weights and biases of the Q-network using gradient descent, Step 8. Repeat Steps 2 - 7 until the maximum number of training rounds is reached;

[0164] where y is the target Q value and γ is the discount factor, Represents the action that maximizes the Q-value at the next moment, Q(s t+1 , a t+1 ) represents the expected reward that can be obtained by executing the action at the state at the next moment. The Q-value is used to represent the expected cumulative reward that can be obtained by taking a certain action (action) under a given state. s t+1 is the state at the next time t + 1, and a t+1 is the action at the next time t + 1;

[0165] Fuzzy control part:

[0166] Construct a fuzzy control rule base: Based on inputs such as position error and force error, output appropriate control signals; if the position error is large and the force error is small, output a "large adjustment" control signal; if the position error is small and the force error is large, output a "fine adjustment" control signal;

[0167] Fuzzy logic reasoning:

[0168] Input variables: position error E p and force error E f , the output variable is the control signal for the adjustment amplitude. Define several fuzzy sets for the input variables and output variables respectively. Each fuzzy set corresponds to a membership function. The fuzzy rules are set as follows:

[0169] If E p is large and E f is small, then a large adjustment amplitude. If E p is small and E f is large, then a small adjustment amplitude;

[0170] Then, according to the membership degrees of the input variables, use fuzzy logic reasoning to obtain the membership degrees of the output variables, and then through defuzzification methods (such as the centroid method), convert the fuzzy sets into control signals for amplitude adjustment;

[0171] Step 9: Actuator control method

[0172] Control signal generation, by combining reinforcement learning and fuzzy control algorithms, generate a comprehensive control signal to ensure the optimal movement path and force of the lifting cylinder; Comprehensive control signal:

[0173] a t = λ·a DQN +(1 - λ)·a Fuzzy

[0174] a t is the comprehensive control signal, λ is the weight coefficient, which adjusts the proportion of the two control signals. a DQN is the control signal generated by reinforcement learning, and a Fuzzy is the control signal generated by fuzzy control;

[0175] Control signal transmission: Transmit the generated control signal to the actuator of the lifting cylinder to achieve real-time adjustment:

[0176] Electric lifting cylinder: Adjust the height according to the input voltage signal;

[0177] Hydraulic lifting cylinder: Adjust the height according to the input hydraulic signal;

[0178] Control signal application: Adjust the movement path and force of the lifting cylinder according to the comprehensive control signal;

[0179] V = k·a t

[0180] V is the voltage signal, k is the proportionality coefficient, and the control signal is converted into a voltage signal;

[0181] P = k2·a t

[0182] P is the hydraulic signal, k2 is the proportionality coefficient, and the control signal is converted into a hydraulic signal;

[0183] Numerical examples are as follows:

[0184] x top = 100.4mm

[0185] x bottom = 100.0mm

[0186] F top = 500N

[0187] F bottom = 495N

[0188] v = 10mm / s

[0189] v set = 12mm / s

[0190] Position error: E p = |100.4 - 100.0| = 0.4

[0191] Force error: E f = |500 - 495| = 5

[0192] Velocity error: E v = |12 - 10| = 2

[0193] Reward function calculation: Assume the weight coefficients α = 1, β = 0.5, γ = 0.2;

[0194] r t=-1·0.4 - 0.5·5 - 0.2·2 = -0.4 - 2.5 - 0.4 = -3.3

[0195] Composite control signal calculation: Assume λ = 0.7, and the control signal generated by reinforcement learning is a DQN = rising, and the control signal generated by fuzzy control is a Fuzzy = large adjustment, then:

[0196] a t = 0.7·rising + 0.3·large adjustment

[0197] 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 perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. Die for outer covering parts of electric vehicle, including a fixed base (1), characterized in that: One side of the top end of the fixed base (1) is fixedly installed with a vertical plate (11), the top end of the vertical plate (11) is fixedly installed with a top plate (12), the middle part of the top plate (12) is fixedly installed with a lifting cylinder (14), the movable end of the lifting cylinder (14) is fixedly installed with a lifting platform (16), the bottom end of the lifting platform (16) is fixedly installed with a top mold (17), the other side of the top end of the fixed base (1) is fixedly installed with two limit plates (2), a lead screw (4) is rotatably connected between the two limit plates (2), the middle part of the lead screw (4) is threadedly connected with a first movable block (3) slidably connected with the fixed base (1), a demolding mechanism (6) is arranged at the top end of the first movable block (3), a bottom mold (7) is fixedly installed at the top end of the demolding mechanism (6), the top mold (17) and the bottom mold (7) are arranged corresponding to each other, a metering mechanism (9) is fixedly installed at the top end of one side of the vertical plate (11), and a heat dissipation mechanism (8) is fixedly installed at the bottom end of one side of the vertical plate (11).

2. The mold for the outer cover of an electric vehicle according to claim 1, characterized in that: The demolding mechanism (6) includes an assembly table (61) and a demolding table (65). A shock absorber (62) is fixedly installed in the middle of the top end of the assembly table (61), and the top end of the shock absorber (62) is fixedly connected with the middle of the bottom end of the demolding table (65). Fixed shells (63) are fixedly installed on both sides of the top end of the assembly table (61). Shock absorption springs (66) are fixedly installed at the top ends of the two fixed shells (63). Push rods (64) are fixedly installed at the top ends of the two shock absorption springs (66). The top ends of the two push rods (64) are respectively fixedly connected with both sides of the bottom end of the demolding table (65).

3. The electric vehicle exterior cover mold according to claim 2, characterized in that: The bottom end of the assembly table (61) is fixedly connected with the top end of the first movable block (3), and the top end of the demolding table (65) is fixedly connected with the bottom end of the bottom mold (7).

4. The mold for the outer cover of an electric vehicle according to claim 1, characterized in that: The heat dissipation mechanism (8) includes an angle seat (83) and an angle plate (84). One end of the angle plate (84) is rotatably connected with the inside of the angle seat (83). An air jet cover (85) is fixedly communicated with the bottom end of the angle plate (84). A movable frame (86) is fixedly installed at the top end of the angle plate (84). A second movable block (88) is slidably connected with the middle part of the movable frame (86). A push cylinder (87) is fixedly installed on one side of the second movable block (88). The fixed end of the push cylinder (87) and one side of the angle seat (83) are both fixedly connected with the side of the vertical plate (11) facing each other.

5. The electric vehicle outer covering part mold according to claim 4, characterized in that: A fan body (81) is fixedly installed at the bottom end of one side of the vertical plate (11). An air inlet hose (82) is fixedly communicated with the air inlet of the fan body (81). An air supply hose (89) extending into the air jet cover (85) is fixedly communicated with the air outlet of the fan body (81).

6. The mold for the outer covering part of the electric vehicle according to claim 1, wherein: The quantitative mechanism (9) includes a quantitative shell (91) and a quantitative frame (96). The top end of the quantitative shell (91) is fixedly connected to the bottom end of the quantitative frame (96). Inside the quantitative shell (91), a heating plate (93) is fixedly installed. The top end of the inner wall of the quantitative shell (91) is slidably connected to a sealing plate (94). The top end of the sealing plate (94) is fixedly installed with a vertical rod (95) extending to the outside. In the middle of the quantitative frame (96), a quantitative cylinder (97) is fixedly installed. The movable end of the quantitative cylinder (97) is fixedly connected to one side facing the vertical rod (95). On the surface of the quantitative shell (91), a number of capacity lines (98) are provided. The bottom end of the quantitative shell (91) is fixedly communicated with a suction head (92). One side of the quantitative frame (96) is fixedly connected to one side facing the vertical plate (11).

7. The mold for the outer covering part of the electric vehicle according to claim 6, characterized in that: On the surface of the vertical plate (11), a pump body (19) is fixedly installed. The liquid inlet of the pump body (19) is fixedly communicated with a suction hose (10) extending into the interior of the quantitative shell (91). The liquid outlet of the pump body (19) is fixedly communicated with a delivery hose (18) extending into the interior of the top mold (17).

8. The mold for the outer covering of an electric vehicle according to claim 1, characterized in that: On both sides of the top end of the lifting table (16), sliding frames (13) slidably connected to the top plate (12) are fixedly installed. Between the two sliding frames (13) and the top plate (12), connecting springs (15) are fixedly installed. On the surface of one of the limiting plates (2), a stepping motor (5) for driving the rotation of the lead screw (4) is fixedly installed.

9. The manufacturing method of the outer covering part mold of an electric vehicle according to any one of claims 1-8, characterized in that, It includes the following steps: Step 1, raw material quantification: The quantitative mechanism (9) quantifies the materials for manufacturing the outer covering parts of electric vehicles. Step 2, mold alignment: The first movable block (3) slides along the lead screw (4) to longitudinally align the bottom mold (7) with the top mold (17). Step 3, mold clamping: The lifting cylinder (14) pushes the lifting table (16) to clamp the top mold (17) inside the bottom mold (7). Step 4, material injection: The pump body (19) extracts the quantified raw materials and injects them between the top mold (17) and the bottom mold (7). Step 5, shaping and heat dissipation: The heat dissipation mechanism (8) extracts low-temperature gas through the fan body (81) and blows out the low-temperature gas through the air jet cover (85) to dissipate heat and shape the product between the molds. Step 6, demolding and packaging: The demolding mechanism (6) pushes the bottom mold (7) from the bottom, peels the product from between the bottom mold (7) and the top mold (17), and then packages it.

10. The manufacturing method of the outer covering part mold of an electric vehicle according to claim 9, characterized in that, In Step 3, introducing an intelligent algorithm to make the top mold (17) accurately clamp inside the bottom mold (7) includes the following steps: Step 1: Sensor installation Use high-precision laser displacement sensors as position sensors. Install position sensors at the four corners of the top mold and the bottom mold to ensure that the position changes of the molds can be accurately monitored. By comparing the sensor readings with the actual positions, necessary deviation adjustments are made. X top = [x top1 , x top2 , x top3 , x top4 ​ x bottom = [x bottom1 , x bottom2 , x bottom3 , x bottom4 ​ Among them, x top refers to the installation position on the top mold, x top1 is the first corner of the fixed mold position, x top2 is the second corner of the fixed mold position, x top3 is the third corner of the fixed mold position, x top4 is the fourth corner of the fixed mold position, x bottom refers to the installation position on the bottom mold, x bottom1 is the first corner of the bottom mold position, x bottom2 is the second corner of the bottom mold position, x bottom3 is the third corner of the bottom mold position, x bottom4 is the fourth corner of the bottom mold position; Average value of the top mold position: Average value of the bottom mold position: Position error: Among them, is the mean value of the top mold position, x topi is the i-th corner of the top mold, is the mean value of the bottom mold position, x bottomi is the i-th corner of the bottom mold, E p is the position error value; Use a high-precision strain gauge force sensor as the force sensor, which is fixed at the top and bottom of the lifting cylinder to ensure that the sensor can accurately sense the force applied during the lifting and lowering of the cylinder. Through range calibration, apply a known force, record the sensor readings, and correct the range coefficient; Force mean: Force error: E f = |F top - F bottom | Among them, is the average value of the forces at the top and bottom, F top is the force at the top, F bottom is the force at the bottom, E f is the error of the force; Use a high-precision linear encoder as the speed sensor, which is installed in the middle of the lifting cylinder to monitor the lifting speed in real time. Compare the sensor readings with the actual speed and make necessary deviation adjustments; Speed error: E v = |v set - v| where E v is the speed error, v set is the set speed, and v is the real-time speed; Step 2: Data preprocessing Select median filtering to ensure the smoothness and accuracy of the data, Among them, median represents the median filtering function, x(t) represents the original data at time t, x(t-N+1) is the data point at the earliest time within the sliding window, and N is the size of the sliding window. is the denoised data at time t; Step 3: Model usage and control prediction The present invention predicts and optimizes the model by using the reinforcement learning algorithm and the fuzzy control algorithm to ensure the fast and stable buckling of the top die and the bottom die. The specific operations are as follows: First, in the simulation environment, use DQN for policy training, combine the fuzzy control rule base for initial optimization, and then regularly evaluate and adjust the model parameters to ensure the effectiveness of the control strategy. After the training is completed, apply the trained strategy to the actual system and perform fine-tuning in combination with fuzzy control; Reinforcement learning DQN part: The parameters of the state space s include position, force, and speed, and the specific definitions are as follows: s = [x top , x bottom , F top , F bottom , v] The action space a includes the control signals of the lifting cylinder: a = [rise, fall, stop] Rise: Increase the height of the lifting cylinder; Fall: Decrease the height of the lifting cylinder; Stop: Maintain the current height; The reward function is set according to the speed, stability, and mechanical wear of the die buckling: Reward function r t : r t = -α·E p -β·E f -γ·E v E p is the position error, and α, β, γ are weight coefficients used to adjust the influence of each error on the total reward, E f is the force error, E v is the velocity error; Deep Q-network DQN training: Initialization: Initialize the weights and biases of the Q-network; Initialize the experience replay buffer; Training process: Step 1. Randomly select an initial state s0 in the simulation environment. Step 2. Select the action a at the current time t , Step 3. Execute the action and observe the reward r at the current time t and the next-time state S t+1 ; Step 4. Store the experience (S t , a t , r t , S t+1 ) into the experience replay buffer; Step 5. Randomly extract a small batch of samples from the experience replay buffer; Step 6. Calculate the target Q value: where y is the target Q value, and γ is the discount factor, represents the action that maximizes the Q value at the next time step, and Q(s t+1 , a t+1 ) represents the expected reward that can be obtained by executing the action at the state at the next time step. The Q value is used to represent the expected cumulative reward that can be obtained by taking a certain action action in a given state. s t+1 is the state at the next time t + 1, and a t+1 is the action at the next time t + 1; Step 7. Update the weights and biases of the Q-network using the gradient descent method; Step 8. Repeat steps 2-7 until the maximum number of training rounds is reached; Fuzzy control part: Construct a fuzzy control rule base: Based on inputs such as position error and force error, output appropriate control signals; If the position error is large and the force error is small, output a "large adjustment" control signal; If the position error is small and the force error is large, output a "fine adjustment" control signal; Fuzzy logic inference: Input variables: position error E p and force error E f , the output variable is the control signal of the adjustment range. A number of fuzzy sets are defined for the input variables and output variables respectively, and each fuzzy set corresponds to a membership function. The fuzzy rules are set as follows: If E p is large and E f is small, then a large adjustment amplitude, if E p is small and E f is large, then a small adjustment amplitude; Then, according to the membership degrees of the input variables, use fuzzy logic inference to obtain the membership degrees of the output variables, and then through the defuzzification method, convert the fuzzy set into a control signal for amplitude adjustment; Step 9: Actuator control method Control signal generation, by combining the reinforcement learning and fuzzy control algorithms, generate a comprehensive control signal to ensure the optimal movement path and force of the lifting cylinder; Comprehensive control signal: a t = λ · a DQN + (1 - λ) · a Fuzzy a t is the comprehensive control signal, λ is the weight coefficient for adjusting the proportion of the two control signals, a DQN is the control signal generated by reinforcement learning, a Fuzzy is the control signal generated by fuzzy control; Control signal transmission, transmit the generated control signal to the actuator of the lifting cylinder to achieve real-time adjustment: 1 Electric lifting cylinder: Adjust the height according to the input voltage signal; Hydraulic lifting cylinder: Adjust the height according to the input hydraulic signal; Control signal application: Adjust the movement path and force of the lifting cylinder according to the comprehensive control signal; V = k·a t V is the voltage signal, k is the proportional coefficient, and the control signal is converted into a voltage signal; P = k2·a t P is the hydraulic signal, k2 is the proportional coefficient, and the control signal is converted into a hydraulic signal.

Images