Double-blade semi-open sewage pump impeller and manufacturing device and method thereof

Abstract

The present invention discloses a double-blade semi-open sewage pump impeller and a manufacturing device and method thereof. The present invention relates to the field of impeller technology, including a workbench, a front end of the workbench fixedly connected to a connecting plate, a surface of the workbench fixedly connected to a fixing frame, a top end of the fixing frame is provided with a pressing mechanism, the left and right ends of the workbench are respectively provided with a movable flipping mechanism and a movable clamping mechanism, and the left and right ends of the connecting plate are respectively provided with a rotating heating treatment mechanism and a cooling mechanism. The advantages of the present invention are: first, the prepared raw materials are poured into the mold cavity, the first driving motor drives the first telescopic rod to extend, thereby pushing out the bottom fixed plate so that the upper and lower membranes can be accurately matched, and a certain time is maintained for the material to cool and solidify, then the first driving motor contracts, the first controller pushes out the bottom plate, and when the rear end of the slide plate moves to the front end of the slide groove, the first rotating motor drives the fixing rod to rotate, thereby pouring the solidified impeller into the heating furnace.

Description

Technical Field

[0001] The present invention relates to the technical field of impellers, in particular to a double-blade semi-open sewage pump impeller and a manufacturing device and method thereof. Background Art

[0002] In the field of sewage treatment engineering, sewage pumps are a common equipment used to transport sewage from one location to another for treatment. The performance of sewage pumps is affected by the impeller design. Traditional sewage pump impellers usually adopt single-blade or multi-blade designs, which have some limitations and problems, such as low fluid flow efficiency and easy clogging.

[0003] After searching, the applicant found that a Chinese patent disclosed "a double-blade semi-open sewage pump impeller and its manufacturing device and method", and its publication (announcement) number is "CN108468654A". The patent mainly includes an impeller cover and a sleeve located on the impeller cover. The sleeve is a cylindrical structure on one side of the impeller cover, and the sleeve is a truncated cone structure on the other side of the impeller cover. An axial hole is processed along the axial centerline direction of the sleeve, and a rotating shaft is installed in the axial hole. A plurality of blades are arranged on the impeller cover. The semi-open impeller manufacturing method includes establishing a semi-open impeller three-dimensional CAD model, model layering and slicing, 3D printing, post-printing processing, and grinding the semi-open impeller blank. However, the above-mentioned prior art uses 3D printing technology to print the designed impeller model layer by layer, which is only suitable for small-batch customized production and cannot be achieved in large-scale production. In addition, the surface quality of the impeller after molding is not good enough. For this reason, we propose a double-blade semi-open sewage pump impeller and its manufacturing device and method. Summary of the Invention

[0004] The object of the present invention is to provide a double-blade semi-open sewage pump impeller and a manufacturing device and method thereof.

[0005] In order to solve the problems raised in the above background technology, the present invention provides the following technical solutions: A double-blade semi-open sewage pump impeller manufacturing device includes a workbench, a connecting plate is fixedly connected to the front end of the workbench, a fixing frame is fixedly connected to the surface of the workbench, a pressing mechanism is provided at the top of the fixing frame, a movable flipping mechanism and a movable clamping mechanism are respectively provided at the left and right ends of the workbench, a rotary heating treatment mechanism and a cooling mechanism are respectively provided at the left and right ends of the connecting plate, the pressing mechanism includes a first drive motor fixedly mounted on the top of the fixing frame, the output end of the first drive motor is fixedly connected to a first telescopic rod, the end of the first telescopic rod away from the first drive motor is fixedly connected to a receiving plate, the four corners of the bottom of the receiving plate are fixedly connected to telescopic columns, the ends of the four telescopic columns away from the receiving plate are commonly fixedly connected to the fixing plate, the bottom surface of the fixing plate is fixedly connected to an upper film, the movable flipping mechanism includes a first telescopic controller, the output end of the first telescopic controller is fixedly connected to a bottom plate via a connecting shaft, a forming plate is provided on the surface of the bottom plate, a cavity is opened on the surface of the forming plate, and a lower film is fixedly mounted on the bottom surface of the cavity.

[0006] As a further solution of the present invention: the circumferential surface of the four telescopic columns is sleeved with a telescopic spring, the upper and lower ends of the telescopic spring are respectively fixedly installed on the bottom surface of the receiving plate and the surface of the fixed plate, and the upper membrane and the lower membrane are on the same vertical line.

[0007] As a further solution of the present invention: a through groove is opened at the right end of the surface of the workbench, a grinding assembly is fixedly installed on the rear side of the right end of the workbench through a motor seat, and a rear plate is fixedly installed on the rear bottom end of the fixing frame.

[0008] As a further solution of the present invention: side panels are fixedly installed on the surface of the base plate and located on the left and right sides of the forming plate, the outer front end of the left side panel is fixedly connected to a first rotating motor, the output end of the first rotating motor is fixedly installed with a fixing rod, and the fixing rod is fixedly installed on the inner front end of the forming plate, and a bearing seat adapted to the fixing rod is fixedly installed on the surface of the right side panel, a slide is fixedly connected to the bottom surface of the base plate, a slide groove adapted to the slide is provided on the surface of the workbench, and the first telescopic controller is fixedly installed on the outer surface of the rear plate.

[0009] As a further solution of the present invention: the rotary heating treatment mechanism includes a heating furnace, a top cover is provided on the top of the heating furnace, and a vacuum exhaust hole is provided at one end of the surface of the top cover, the heating furnace is fixedly installed at the inner left end of the connecting plate, a heating coil is fixedly installed on the inner wall of the heating furnace, a temperature display controller is fixedly installed on the front outer surface of the heating furnace, a second rotating motor is fixedly connected to the bottom outer surface of the heating furnace, a rotating rod is fixedly connected to the output end of the second rotating motor, and a receiving metal disk is fixedly connected to the end of the rotating rod away from the second rotating motor.

[0010] As a further solution of the present invention: the cooling mechanism includes a cooling box fixedly installed inside the right end of the connecting plate, the right side surface of the cooling box is fixedly connected to the box body, a circulating pump is provided on the top of the box body, and a water inlet pipe is fixedly installed on the output end of the circulating pump, one end of the water inlet pipe is arranged inside the box body, and the other end of the water inlet pipe is fixedly connected to a condenser, and the end of the condenser away from the water inlet pipe is fixedly connected to a return water pipe, and the end of the return water pipe away from the condenser is arranged in the box body, the condenser is fixedly installed on the inner wall of the cooling box, and the condenser is arranged in a spiral descending manner.

[0011] As a further solution of the present invention: the mobile clamping mechanism includes a limiting groove opened at the right end of the workbench surface, and there are two limiting grooves, and a plurality of sliders are slidably connected in the two limiting grooves, and two mounting plates are fixedly connected to the surfaces of the plurality of sliders, and a second driving motor is fixedly installed between the two mounting plates through a driving motor seat, and a telescopic push rod is provided inside the limiting groove, one end of the telescopic push rod is fixedly connected to the slider, and the output end of the second driving motor is fixedly connected to the second telescopic rod, and the end of the second telescopic rod away from the second driving motor is fixedly connected to the second telescopic controller, and the upper and lower ends of the second telescopic controller are provided with clamping plates, and the surface of the clamping plate is provided with an anti-slip pad.

[0012] In addition, the present invention also provides a method for manufacturing a double-blade semi-open sewage pump impeller, the specific steps of which are as follows:

[0013] Step 1: First, draw a detailed drawing of the double-blade semi-open sewage pump impeller, make a mold suitable for the shape of the impeller, then introduce the material for making the impeller into the mold cavity, and the first drive motor drives the first telescopic rod to extend, thereby pushing out the bottom fixed plate, so that the upper and lower membranes can accurately match, and keep a certain time for the material to cool and solidify to ensure that the molded product has sufficient hardness and strength. Then the first drive motor contracts, and under the action of the first telescopic controller, the bottom plate is pushed out. When the rear end of the slide moves to the front end of the slide, the first rotary motor drives the fixed rod to rotate, thereby realizing the flipping of the molding plate, making it convenient to pour the solidified impeller onto the receiving metal plate in the heating furnace;

[0014] Step 2: When the solidified impeller is poured onto the receiving metal plate in the heating furnace, the top cover of the heating furnace is closed and the interior of the heating furnace is vacuumed by external equipment to prevent the impeller surface from being oxidized. Then the second rotating motor drives the rotating rod to rotate so that the impeller is heated more evenly. Then the impeller is placed in the cooling box and the circulating pump is immediately turned on. The coolant circulates in the condenser tube to achieve a cooling effect.

[0015] Step 3: The cooled impeller is placed between the two clamping plates. The second telescopic controller contracts the clamping plates to complete the clamping of the impeller. The second telescopic rod is then pushed out by the second drive motor to facilitate the polishing of the impeller surface through the polishing assembly to ensure that the impeller surface is flat and smooth.

[0016] In addition, the present invention also provides a double-blade semi-open sewage pump impeller, including a double-blade semi-open impeller body, the double-blade semi-open impeller body including an impeller cover body, a shaft sleeve is provided through the surface axis of the impeller cover body, an axial hole is opened at the surface axis of the shaft sleeve, and a plurality of blades are provided on the impeller cover body.

[0017] By adopting the above technical solution, compared with the prior art, the beneficial effects of the present invention are:

[0018] 1. The present invention provides a pressing mechanism and a movable flipping mechanism. First, the prepared raw material is poured into the mold cavity. The first drive motor drives the first telescopic rod to extend, thereby pushing out the bottom fixed plate, so that the upper and lower films can accurately match. A certain period of time is maintained to allow the material to cool and solidify. Then, the first drive motor contracts. Under the action of the first telescopic controller, the bottom plate is pushed out. When the rear end of the slide moves to the front end of the chute, the first rotary motor drives the fixed rod to rotate, thereby realizing the rotation of the forming plate. The solidified impeller is poured onto the receiving metal plate in the heating furnace, avoiding burns caused by manual contact with the finished product.

[0019] 2. The present invention is provided with a rotating heating treatment mechanism and a cooling mechanism. When the solidified impeller is poured onto the receiving metal plate in the heating furnace, the top cover is closed and the interior of the heating furnace is evacuated by an external device to prevent oxidation of the impeller surface. Then, the second rotating motor drives the rotating rod to rotate, so that the impeller is heated more evenly. The impeller is kept at the solution temperature for a certain period of time to ensure that the alloy elements in the grains are fully dissolved and the internal stress is eliminated. Then, the impeller is placed in a cooling box, and the circulating pump is turned on. The coolant circulates in the cooling pipe to achieve a rapid cooling effect. In this way, a high-hardness martensite structure can be formed on the impeller surface, thereby improving its wear resistance.

[0020] 3. The present invention sets a mobile clamping mechanism, and the cooled impeller is placed between two clamping plates. The second telescopic controller contracts the clamping plates to complete the clamping of the impeller, and then the second telescopic rod is pushed out by the second drive motor, so that the impeller surface can be polished by the polishing assembly to ensure that the impeller surface is flat and smooth.

[0021] 4. Improve the automation and intelligence level of the manufacturing process

[0022] Through real-time monitoring and adjustments by the intelligent controller, the manufacturing process is fully automated, eliminating the need for human intervention. The intelligent controller automatically adjusts the drive motor speed, telescopic rod position, and forming plate angle to ensure optimal operation at every manufacturing step. Intelligent algorithms enable the controller to dynamically adjust control strategies based on real-time data to adapt to varying production conditions and requirements. Through experience replay and online learning, the intelligent controller continuously optimizes its strategies, enhancing the intelligence of the manufacturing process.

[0023] 5. Improve the precision and quality of the manufacturing process

[0024] By monitoring key state variables (such as material temperature, cooling time, and curing hardness) in real time, the intelligent controller precisely adjusts each control parameter to ensure optimal performance at every step of the manufacturing process. During the material introduction and cooling and curing stages, the intelligent controller precisely controls temperature and time to ensure uniform material introduction and optimal curing. Through intelligent algorithm optimization, the resulting impellers are more consistently high quality and have a reduced defective rate. The controller can rapidly adjust operations based on real-time feedback, reducing errors and defects in the manufacturing process.

[0025] 6. Improve production efficiency and reduce costs

[0026] Intelligent controllers can optimize control strategies, reducing time at each production stage and improving overall production efficiency. By optimizing cooling system parameters, cooling time can be minimized while ensuring curing quality. By reducing defective product rates and improving production efficiency, intelligent controllers can significantly lower production costs. Intelligent algorithms can optimize energy consumption, enabling equipment to operate efficiently while reducing energy consumption.

[0027] 7. Improve the robustness and adaptability of the system

[0028] Through continuous online learning and optimization, the intelligent controller adapts to the various uncertainties and changes that arise during the manufacturing process, ensuring stable system operation. When faced with fluctuations in material properties or changes in environmental conditions, the controller can quickly adjust its strategy to ensure production quality and efficiency. Intelligent algorithms enable the controller to continuously accumulate and learn from actual operational data, gradually optimizing control strategies and improving the system's adaptability to varying production conditions and requirements. The controller can dynamically adjust control parameters based on historical and real-time data to adapt to a variety of manufacturing scenarios.

[0029] 8. Achieve visualization and transparency of the manufacturing process

[0030] Real-time monitoring of system status and operating data, displayed through charts and reports, allows operators to intuitively understand the production process and equipment operating conditions. Material temperature, drive motor position, and forming plate angle are monitored in real time, and this data is visualized for analysis and decision-making. By recording and analyzing every step of the production process in detail, the entire manufacturing process is traceable and transparent, facilitating quality management and optimization improvements. By recording the manufacturing data of each impeller, any step in the production process can be traced back to identify the root cause of problems and implement improvements. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a three-dimensional schematic diagram of an embodiment of the present invention;

[0032] Figure 2 This is a schematic top view of an embodiment of the present invention;

[0033] Figure 3 This is a front plan view schematic diagram of an embodiment of the present invention;

[0034] Figure 4 Schematic diagram of the mobile flipping mechanism in an embodiment of the present invention;

[0035] Figure 5 This is a schematic front view of a partial cross-section of an embodiment of the present invention;

[0036] Figure 6 Schematic diagram of the pressing mechanism in an embodiment of the present invention;

[0037] Figure 7 Schematic diagram of a mobile clamping mechanism in an embodiment of the present invention;

[0038] Figure 8 Schematic diagram of a double-blade semi-open impeller body in an embodiment of the present invention.

[0039] Figure 9 This is the flow chart of the intelligent algorithm of step 1 of the present invention

[0040] In the figure: 1. workbench; 2. fixed frame; 3. connecting plate; 4. pressing mechanism; 5. mobile turning mechanism; 6. rotary heating treatment mechanism; 7. cooling mechanism; 8. mobile clamping mechanism; 9. grinding assembly; 10. through slot; 11. back plate; 12. upper film; 13. lower film; 14. cavity; 41. first driving motor; 42. first telescopic rod; 43. receiving plate; 44. telescopic column; 45. telescopic spring; 46. fixed plate; 51. first telescopic controller; 52. slide plate; 53. first rotating motor; 54. chute; 55. bottom plate; 56. forming plate; 57. fixed rod; 58. side Plate; 61. Heating furnace; 62. Vacuum exhaust hole; 63. Second rotating motor; 64. Temperature display controller; 65. Rotating rod; 66. Receiving metal plate; 67. Heating coil; 71. Cooling box; 72. Box body; 73. Circulation pump; 74. Condenser; 75. Water inlet pipe; 76. Return pipe; 81. Limiting groove; 82. Telescopic push rod; 83. Mounting plate; 84. Second driving motor; 85. Second telescopic rod; 86. Second telescopic controller; 87. Clamping plate; 15. Double-blade semi-open impeller body; 151. Impeller cover; 152. Shaft sleeve; 153. Shaft hole; 154. Blade. DETAILED DESCRIPTION

[0041] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0042] In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0043] Please see the attached Figure 1 -Attached Figure 8The present invention relates to a manufacturing device for a double-blade semi-open sewage pump impeller, comprising a workbench 1, a connecting plate 3 being fixedly connected to the front end of the workbench 1, a fixing frame 2 being fixedly connected to the surface of the workbench 1, a pressing mechanism 4 being provided at the top of the fixing frame 2, a moving flipping mechanism 5 and a moving clamping mechanism 8 being provided at the left and right ends of the workbench 1, a rotating heating treatment mechanism 6 and a cooling mechanism 7 being provided at the left and right ends of the connecting plate 3, the pressing mechanism 4 comprising a first driving motor 41 fixedly mounted on the top of the fixing frame 2, the output end of the first driving motor 41 being fixedly connected to a first telescopic rod 42, and the first telescopic rod 43 being provided at the top end of the fixing frame 2. The end of the retracting rod 42 away from the first driving motor 41 is fixedly connected to the receiving plate 43, and the four corners of the bottom of the receiving plate 43 are fixedly connected to the telescopic columns 44. The ends of the four telescopic columns 44 away from the receiving plate 43 are commonly fixedly connected to the fixed plate 46. The bottom surface of the fixed plate 46 is fixedly connected to the upper membrane 12. The mobile flip mechanism 5 includes a first telescopic controller 51. The output end of the first telescopic controller 51 is fixedly connected to the bottom plate 55 through a connecting shaft. The surface of the bottom plate 55 is provided with a forming plate 56. The surface of the forming plate 56 is provided with a cavity 14. The bottom surface of the cavity 14 is fixedly mounted with the lower membrane 13.

[0044] In one embodiment of the present invention: the circumferential surface of the four telescopic columns 44 is sleeved with a telescopic spring 45, and the upper and lower ends of the telescopic spring 45 are fixedly installed on the bottom surface of the receiving plate 43 and the surface of the fixed plate 46 respectively, and the upper membrane 12 and the lower membrane 13 are on the same vertical line.

[0045] In one embodiment of the present invention: a through groove 10 is opened at the right end of the surface of the workbench 1, a grinding assembly 9 is fixedly installed on the rear side of the right end of the workbench 1 through a motor seat, and a rear plate 11 is fixedly installed on the rear bottom end of the fixing frame 2.

[0046] In one embodiment of the present invention: side panels 58 are fixedly installed on the surface of the bottom plate 55 and located on the left and right sides of the forming plate 56, the outer front end of the left side panel 58 is fixedly connected to the first rotating motor 53, the output end of the first rotating motor 53 is fixedly installed with a fixing rod 57, the fixing rod 57 is fixedly installed on the inner front end of the forming plate 56, the surface of the right side panel 58 is fixedly installed with a bearing seat adapted to the fixing rod 57, the bottom surface of the bottom plate 55 is fixedly connected to the slide 52, the surface of the workbench 1 is provided with a slide groove 54 adapted to the slide 52, and the first telescopic controller 51 is fixedly installed on the outer surface of the rear plate 11.

[0047] In one embodiment of the present invention: the rotary heating treatment mechanism 6 includes a heating furnace 61, a top cover is provided on the top of the heating furnace 61, and a vacuum exhaust hole 62 is provided at one end of the surface of the top cover, the heating furnace 61 is fixedly installed at the inner left end of the connecting plate 3, a heating coil 67 is fixedly installed on the inner wall of the heating furnace 61, a temperature display controller 64 is fixedly installed on the front outer surface of the heating furnace 61, a second rotating motor 63 is fixedly connected to the bottom outer surface of the heating furnace 61, the output end of the second rotating motor 63 is fixedly connected to a rotating rod 65, and the end of the rotating rod 65 away from the second rotating motor 63 is fixedly connected to a receiving metal plate 66.

[0048] In one embodiment of the present invention: the cooling mechanism 7 includes a cooling box 71 fixedly installed inside the right end of the connecting plate 3, the right side surface of the cooling box 71 is fixedly connected to the box body 72, the top of the box body 72 is provided with a circulating pump 73, the output end of the circulating pump 73 is fixedly installed with a water inlet pipe 75, one end of the water inlet pipe 75 is arranged inside the box body 72, the other end of the water inlet pipe 75 is fixedly connected to a condenser 74, the end of the condenser 74 away from the water inlet pipe 75 is fixedly connected to a return pipe 76, the end of the return pipe 76 away from the condenser 74 is arranged in the box body 72, the condenser 74 is fixedly installed on the inner wall of the cooling box 71, and the condenser 74 is arranged in a spiral descending manner.

[0049] In one embodiment of the present invention: the mobile clamping mechanism 8 includes a limiting groove 81 opened at the right end of the surface of the workbench 1, and there are two limiting grooves 81. A plurality of sliders are slidably connected inside the two limiting grooves 81, and two mounting plates 83 are fixedly connected to the surfaces of the plurality of sliders. A second drive motor 84 is fixedly installed between the two mounting plates 83 through a drive motor seat. A telescopic push rod 82 is provided inside the limiting groove 81, and one end of the telescopic push rod 82 is fixedly connected to the slider. The output end of the second drive motor 84 is fixedly connected to a second telescopic rod 85, and the end of the second telescopic rod 85 away from the second drive motor 84 is fixedly connected to a second telescopic controller 86. The upper and lower ends of the second telescopic controller 86 are provided with clamping plates 87, and the surface of the clamping plate 87 is provided with an anti-slip pad.

[0050] Example 1, please refer to the attached Figure 1 -Attached Figure 8 The circumferential surface of the four telescopic columns 44 is sleeved with a telescopic spring 45. The upper and lower ends of the telescopic spring 45 are fixedly mounted on the bottom surface of the receiving plate 43 and the surface of the fixed plate 46 respectively. The arrangement of the telescopic spring 45 and the telescopic column 44 ensures that after the fixed plate 46 descends, the upper membrane 12 can be cushioned when squeezing the raw material, thereby preventing deformation caused by extrusion and causing the impeller to be substandard.

[0051] The upper film 12 and the lower film 13 are on the same vertical line. When the upper film 12 and the lower film 13 are on the same vertical line, the upper film 12 and the lower film 13 can be accurately matched, ensuring that the material can be fully filled after being pressed by the pressing mechanism 4 during molding;

[0052] A bearing seat compatible with the fixed rod 57 is fixedly installed on the surface of the right side plate 58, and a slide 52 is fixedly connected to the bottom surface of the base plate 55. A slide groove 54 compatible with the slide 52 is provided on the surface of the workbench 1. The slide 52 moves in the slide groove 54. When the rear end of the slide 52 moves to the front end of the slide groove 54, the fixed rod 57 drives the forming plate 56 to rotate, and the solidified impeller in the cavity 14 is poured into the heating furnace 61.

[0053] Example 2, please refer to the attached Figure 1 -Attached Figure 8 A heating coil 67 is fixedly installed on the inner wall of the heating furnace 61, and a temperature display controller 64 is fixedly installed on the front outer surface of the heating furnace 61. The temperature display controller 64 and the heating coil 67 are electrically connected. The temperature display controller 64 can control and display the temperature in the heating furnace 61. Once the impeller reaches the solution temperature, it needs to be maintained for a period of time to ensure complete dissolution of the grains. The insulation time depends on the material and thickness of the impeller. After heating to the solution temperature, the heating should be stopped in time to avoid excessive heating and causing impeller burning.

[0054] Example 3, please refer to the attached Figure 1 -Attached Figure 8 A through slot 10 is provided on the right end of the surface of the workbench 1. A grinding assembly 9 is fixedly installed on the rear side of the right end of the workbench 1 through the motor seat. The through slot 10 can collect the grinding debris for unified processing;

[0055] One end of the second telescopic rod 85 away from the second drive motor 84 is fixedly connected to the second telescopic controller 86. The upper and lower ends of the second telescopic controller 86 are provided with clamping plates 87. The surface of the clamping plate 87 is provided with an anti-slip pad. The second telescopic controller 86 can control the extension and retraction of the clamping plate 87 to facilitate clamping the impeller. The setting of the anti-slip pad makes the clamping more stable.

[0056] In addition, the present invention also provides a method for manufacturing a double-blade semi-open sewage pump impeller, comprising the following steps:

[0057] Step 1: First, draw a detailed drawing of the double-blade semi-open sewage pump impeller, make a mold suitable for the shape of the impeller, then introduce the material for making the impeller into the mold cavity 14, and the first drive motor 41 drives the first telescopic rod 42 to extend, thereby pushing out the bottom fixed plate 46, so that the upper film 12 and the lower film 13 can be accurately matched, and keep a certain time for the material to cool and solidify to ensure that the molded product has sufficient hardness and strength. Then, the first drive motor 41 contracts, and under the action of the first telescopic controller 51, the bottom plate 55 is pushed out. When the rear end of the slide plate 52 moves to the front end of the slide 54, the first rotating motor 53 drives the fixing rod 57 to rotate, thereby realizing the flipping of the molding plate 56, so that the solidified impeller can be poured onto the receiving metal plate 66 in the heating furnace 61;

[0058] Step 2: When the solidified impeller is poured onto the receiving metal plate 66 in the heating furnace 61, the top cover of the heating furnace 61 is closed and the interior of the heating furnace 61 is evacuated by external equipment to prevent the impeller surface from being oxidized. Then, the second rotating motor 63 drives the rotating rod 65 to rotate so that the impeller is heated more evenly. Then, the impeller is placed in the cooling box 71 and the circulating pump 73 is immediately turned on to circulate the coolant in the condenser pipe 74.

[0059] Step 3: The cooled impeller is placed between the two clamping plates 87. The second telescopic controller 86 contracts the clamping plates 87 to complete the clamping of the impeller. Then, the second drive motor 84 drives the second telescopic rod 85 to be pushed out, so that the impeller surface can be polished by the polishing component 9 to ensure that the impeller surface is flat and smooth.

[0060] In addition, the present invention also provides a double-blade semi-open sewage pump impeller: including a double-blade semi-open impeller body 15, the double-blade semi-open impeller body 15 includes an impeller cover body 151, a shaft sleeve 152 is provided at the surface axis center of the impeller cover body 151, an axis hole 153 is opened at the surface axis center of the shaft sleeve 152, and a plurality of blades 154 are provided on the impeller cover body 151.

[0061] Example 4, please refer to the attached Figure 1 -Attached Figure 8 The coolant circulates in the condenser tube 74 to achieve rapid cooling, thereby forming a high-hardness martensite structure on the impeller surface and improving its wear resistance.

[0062] The second rotating motor 63 drives the rotating rod 65 to rotate, so that the impeller is heated more evenly and the impeller is kept at the solution temperature for a certain period of time to ensure that the alloy elements in the grains are fully dissolved and the internal stress is eliminated.

[0063] Specifically, the material for manufacturing the impeller is introduced into the mold cavity 14, and the first drive motor 41 drives the first telescopic rod 42 to extend, thereby pushing out the fixed plate 46 at the bottom, so that the upper membrane 12 and the lower membrane 13 can be accurately matched. A certain time is maintained to allow the material to cool and solidify. Then the first drive motor 41 contracts, and under the action of the first telescopic controller 51, the bottom plate 55 is pushed out. When the rear end of the slide plate 52 moves to the front end of the slide groove 54, the first rotating motor 53 drives the fixed rod 57 to rotate, thereby realizing the flipping of the molding plate 56, which facilitates pouring the solidified impeller onto the receiving metal plate 66 in the heating furnace 61.

[0064] Specifically, when the solidified impeller is poured onto the receiving metal plate 66 in the heating furnace 61, the top cover of the heating furnace 61 is covered, and the interior of the heating furnace 61 is vacuumed by external equipment to prevent the impeller surface from being oxidized. Then the second rotating motor 63 drives the rotating rod 65 to rotate, so that the impeller is heated more evenly. Then the impeller is placed in the cooling box 71, and the circulating pump 73 is immediately turned on. The coolant circulates in the condenser 74, which has a rapid cooling effect.

[0065] Specifically, the cooled impeller is placed between two clamping plates 87, and the second telescopic controller 86 contracts the clamping plates 87 to complete the clamping of the impeller. Then, the second telescopic rod 85 is pushed out by the second drive motor 84, so that the impeller surface can be polished by the polishing assembly 9 to ensure that the impeller surface is flat and smooth.

[0066] In the above step 1, when manufacturing the double-blade semi-open sewage pump impeller, an intelligent algorithm combining the deep deterministic policy gradient algorithm and the dual deep Q network is introduced to optimize the manufacturing process of the double-blade semi-open sewage pump impeller. This is used to improve the accuracy and quality of the manufacturing process, enhance the automation and intelligence level of the manufacturing process, improve production efficiency and reduce costs, and realize the visualization and transparency of the manufacturing process. The specific process is as follows:

[0067] Step 1: Create an environment model

[0068] Creating a simulation environment

[0069] Simulation goal: To simulate all stages of the manufacturing process for a two-blade semi-open sewage pump impeller, including material introduction, cooling and solidification, and forming plate flipping;

[0070] Simulation tools: Use Python and the SimPy library to create a simulation environment. Use Python and the SimPy library to create a simulation environment, simulate the entire manufacturing process, and collect data on different state variables. SimPy is a Python library for event-driven simulation, suitable for creating discrete event simulation models of manufacturing processes.

[0071] Simulation steps:

[0072] Material import: simulate the process of injecting material into the mold;

[0073] Cooling and solidification: simulate the cooling and solidification process of materials;

[0074] Forming plate flipping: simulates the process of flipping the forming plate to release the curing impeller;

[0075] State variables

[0076] Material temperature (T m ): Indicates the current temperature of the material, ranging from room temperature to curing temperature;

[0077] Cooling time (t cool ): Indicates the material cooling time, set the preset range

[0078] Curing hardness (H cure ): Indicates the curing hardness of the material, ranging from 0 (uncured) to 1 (fully cured);

[0079] Drive motor position (P motor ): Indicates the position of the drive motor, ranging from 0 (starting position) to 1 (ending position);

[0080] Telescopic rod position (P rod ): indicates the position of the telescopic rod, ranging from 0 (retracted) to 1 (fully extended);

[0081] Forming plate angle (A plate ): Indicates the angle of the forming plate, ranging from 0 degrees (horizontal) to 90 degrees (vertical);

[0082] Using deep neural networks to represent environmental dynamics to monitor and control state variables in the impeller manufacturing process;

[0083] Deep neural network model structure: Set up input layers, hidden layers, and output layers in the deep neural network model to model and predict various state variables in the manufacturing process;

[0084] Input layer: includes all state variables (material temperature T m , cooling time t cool , cured hardness H cure , drive motor position P motor , telescopic rod position P rod , forming plate angle A plate );

[0085] Hidden layer: two hidden layers, each containing 64 neurons, with ReLU activation function;

[0086] Output layer: predict all state variables at the next moment;

[0087] Training a deep neural network model:

[0088] Data Collection: Historical and experimental data were collected. The dataset included all state variables in each manufacturing process, recording the entire process from material introduction to forming plate flipping. The dataset also included material temperature, cooling time, curing hardness, drive motor position, telescopic rod position, and forming plate angle. The experimental and historical data were used to train the deep neural network model.

[0089] Optimize control parameters at each stage of the manufacturing process to improve manufacturing efficiency and quality. Use optimizers and loss functions: Use the Adam optimizer and mean square error (MSE) loss function to optimize the deep neural network model to ensure that the model can accurately predict and simulate state changes in the manufacturing process. Adam optimizer formula:

[0090] m t =β1m t-1 +(1-β1)g t

[0091]

[0092]

[0093]

[0094]

[0095] m t : The first moment estimate (momentum) at time t; this is the momentum at the current moment, combined with the momentum m at the previous moment t-1 and the current gradient g t ;

[0096] v t : Second-order moment estimate (square of gradient) at time t; this is the square of the gradient at the current moment, combined with the second-order moment estimate v at the previous moment t+1 and the square of the current gradient

[0097] Momentum m t Bias-corrected estimate of the momentum m t The estimation was performed after bias correction, taking into account the exponential decay of β1;

[0098] Second moment estimate v t The bias-corrected estimate of the second-order moment estimate v t The bias-corrected estimates take into account the exponential decay of β2;

[0099] θt : parameter value at time t; is the updated parameter value, combining the correction estimate of momentum and second-order moment, using learning rate α and a small constant ∈ to prevent division by zero error;

[0100] g t : Gradient at time t This is the gradient value at the current moment;

[0101] β1: exponential decay rate of the first-order moment estimate;

[0102] β2: exponential decay rate of the second-order moment estimate;

[0103] α: learning rate; this is the coefficient used to adjust the parameter update step size;

[0104] ∈: A constant to prevent division by zero errors; usually set to a very small value

[0105] The loss function is the mean square error (MSE), which is used to minimize the error between the predicted value and the true value to ensure the accuracy of the model prediction;

[0106]

[0107] y i : true value;

[0108] Predicted value;

[0109] n: number of samples;

[0110] Step 2. Design an intelligent controller

[0111] Design of intelligent controller based on dual deep Q network algorithm

[0112] The dual-depth Q-network algorithm combines a policy network and a value network to determine the optimal action and evaluate the value of the action, respectively. An intelligent controller was designed using the dual-depth Q-network algorithm to optimize the control parameters of the impeller manufacturing process for a two-blade semi-open sewage pump.

[0113] 1. Strategy network: Receives the current state input of the manufacturing process of the double-blade semi-open sewage pump impeller and outputs corresponding control actions. These actions are used to adjust various control parameters in the manufacturing process to achieve the best manufacturing results;

[0114] Input: Current state s, including the following state variables: Material temperature T m , cooling time t cool , cured hardness H cure , drive motor position P motor , telescopic rod position P rod , forming plate angle A plate ;

[0115] structure:

[0116] Input layer: 6 state variables;

[0117] Hidden layer: two hidden layers, 64 neurons in each layer, and the activation function is ReLU;

[0118] Output layer: controls the action, the activation function is tanh;

[0119] 2. Value Network: Evaluates the value of performing specific actions in the current state of the double-blade semi-open sewage pump impeller manufacturing process, thereby guiding the policy network to select the optimal action in the double-blade semi-open sewage pump impeller manufacturing process;

[0120] Input: A combination of state s and action a, specifically including:

[0121] State s: The same 6 state variables as the policy network input material temperature T m , cooling time t cool , cured hardness H cure , drive motor position P motor , telescopic rod position P rod , forming plate angle A plate ;

[0122] Action a: The control action output by the strategy network, which adjusts the motor speed and telescopic rod position during the manufacturing process of the double-blade semi-open sewage pump impeller;

[0123] Output: Q value, which represents the value of a given state and action in the manufacturing process of a double-blade semi-open sewage pump impeller. The Q value is used to evaluate the value of the current action to help the policy network optimize its output.

[0124] structure:

[0125] Input layer: a combination of state and action;

[0126] Hidden layer: two hidden layers, 64 neurons in each layer, and the activation function is ReLU;

[0127] Output layer: Q value;

[0128] Optimizing Q-value estimation using a dual deep Q-network algorithm

[0129] 3. The dual deep Q network reduces the bias in Q-value estimation by using two independent networks to calculate action selection and action value respectively. During the impeller manufacturing process, immediate rewards can be set based on indicators such as molding quality, cooling rate, and energy consumption. When the material is introduced evenly, the cooling time is appropriate, and the solidification hardness meets the requirements, high rewards are given; otherwise, low rewards or penalties are given.

[0130] The discount factor in the calculation is used to weigh the importance of immediate rewards and future rewards. In the manufacturing process, it is necessary to weigh the current control action and the future molding quality.

[0131] Target Q-network (Q′): An independent target Q-network is used to calculate the target Q-value to reduce estimation bias. The target Q-values for different state and action combinations are calculated through the target Q-network, thereby guiding the policy network to optimize the control action.

[0132] The state and action at the next moment During the manufacturing process, the state and corresponding control action at the next moment are predicted through the current control action and state change; the material temperature and cooling hardness at the next moment are predicted through the current material temperature and cooling time; the drive motor position and telescopic rod position at the next moment are predicted through the current drive motor position and telescopic rod position

[0133] The target value calculation formula of the dual deep Q network is:

[0134] y=rγQ`(s`,argmax a 、Q(s`,a、|θ Q )|θ Q )

[0135] in:

[0136] y: target value, which is the Q-value update target, indicating the value estimation target of the current state-action pair; r: immediate reward, which is the reward obtained after performing the action in the current state;

[0137] Y: Discount factor, used to discount the importance of future rewards, ranging from 0 to 1,

[0138] Q′: target Q network, an independent Q network used to calculate the target value to reduce estimation bias;

[0139] s`: The state at the next moment, the state to be transferred to after executing the current action;

[0140] argmax a `Q(s,a`|θ Q )|θQ): The strategy for selecting an action is to select the action a′ that maximizes the Q value in the next state s′; where Q(s′, a′|θ Q )|θ Q `) represents the Q value calculated using the current Q network;

[0141] θ Q : Parameters of the current Q network, parameters of the current Q network, used to calculate the Q value;

[0142] θ Q`: Parameters of the target Q network, used to calculate the target Q value;

[0143] Step 3: Train the Smart Controller

[0144] 1. Initialize the parameters of the deep deterministic policy gradient and dual deep Q network algorithms

[0145] Parameter initialization: Before training begins, the weights of the policy network and the value network are randomly initialized to break the symmetry and make the network learn different features;

[0146] The target network and Q′ copy the initial weights of the policy network and value network: The target network is used to stabilize the training process. Its initial weights are directly copied from the policy network and value network. As training progresses, the weights of the target network will be slowly updated;

[0147] Set the learning rate α and discount factor γ: Learning rate α: controls the step size of parameter updates to ensure stable and gradual learning; Discount factor γ: weighs the importance of immediate rewards and future rewards;

[0148] 2. Training in a simulation environment

[0149] Training steps

[0150] Simulation run: Start the run in the simulation environment and set the initial state s; the simulation environment simulates all stages of the real manufacturing process, including material introduction, cooling and solidification, and forming plate flipping;

[0151] Action selection: At each time step, the intelligent controller selects an action a based on the current state s. This action is generated by the policy network, which outputs the control parameters motor speed and telescopic rod position adjustment.

[0152] Execute action: Execute the selected action a, causing the manufacturing equipment to operate according to the control parameters and obtain the next state s` and immediate reward r; the immediate reward is based on the effect of the current action, material introduction uniformity, cooling effect and molding quality;

[0153] Experience replay: Use the experience replay mechanism to store state transitions (s, a, r, s`); the experience replay buffer stores these state transitions so that small batches of data can be extracted from them for network updates later; this mechanism can break data correlation and enhance training effects;

[0154] 3. Network Update

[0155] Value network loss function: Calculate the loss function of the value network to optimize the parameters of the value network;

[0156] Loss function:

[0157] L(θ Q)=E s，a，r，s` [(Q(s,a|θ Q )-y) 2 ]

[0158] L(θ Q ): The loss function of the value network, which is used to measure the error between the predicted value of the value network Q and the target value y;

[0159] θ Q : Parameters of the value network, used to represent the weights and biases of the value network. These parameters need to be optimized through training;

[0160] E s，a，r，s `: Expectation, which represents the expected value of all states, actions, rewards and next states;

[0161] Q(s,a|θ Q ): The output of the value network, the estimated value of the value network Q under a given state s and action a, which represents the expected return in state s after taking action a;

[0162] y: target value, the target value to be learned by the value network, used to update the parameters of the value network;

[0163] Policy network gradient: Calculate the gradient of the policy network to optimize the parameters of the policy network;

[0164] Gradient formula:

[0165]

[0166] Policy network gradient, which represents the objective function J versus the policy network parameter θ μ The gradient of is used to update the parameters of the policy network;

[0167] E s : Expectation, which represents the expected value of all states s;

[0168] The gradient of the value network to the action represents the gradient of the value network Q to the action a given state s and action a;

[0169] a = μ(s|θ): the action generated by the policy network, which represents the action a generated by the policy network in a given state s;

[0170] The gradient of the policy network to the parameters indicates the gradient of the policy network to the parameters under a given state s;

[0171] Step 4: Real-time Control and Adjustment

[0172] Use trained intelligent controller

[0173] Deploy intelligent controllers: Ensure that the intelligent controller model has been fully trained and validated in a simulation environment. Through continuous training and testing, the intelligent controller can stably generate reasonable control actions in the simulation environment. Performance evaluation is performed in different simulation scenarios to ensure excellent performance under various operating conditions.

[0174] Deploy the trained model to the actual manufacturing equipment control system. Connect the intelligent controller model to the hardware interface of the actual equipment to ensure accurate transmission of control signals. Integrate the model into the existing manufacturing equipment control system to enable it to take over control of the drive motor, telescopic rod, and forming plate.

[0175] The controller generates control actions based on real-time monitored state variables such as material temperature, telescopic rod position, and forming plate angle. These control actions directly impact the equipment, ensuring optimal operation and ensuring efficient and precise manufacturing.

[0176] Sensors installed at various locations monitor the system status in real time. Based on this real-time status, the intelligent controller selects the optimal control action and adjusts manufacturing process parameters. The real-time data collected by the sensors is transmitted to the control system via a data bus. The intelligent controller analyzes this data in real time to determine whether adjustments are needed.

[0177] The intelligent controller selects the optimal control action based on the real-time status and calculates the optimal motor speed, telescopic rod position and forming plate angle. The controller immediately executes these actions to ensure a stable and efficient manufacturing process.

[0178] Optimize strategies using experience replay

[0179] The intelligent controller continues to use the experience replay mechanism to continuously optimize and update the control strategy. State transition data collected during real-time operation is stored in the experience replay buffer. The intelligent controller is further trained and optimized by collecting state transition data from actual operation. State transition data includes state s, action a, reward r, and next state s`. This data is stored in the experience replay buffer.

[0180] Small batches of data are randomly sampled from the buffer for training. Training is performed periodically, and the sampled data is used to update the policy network and value network to further optimize the control strategy. Data collected from actual operations is typically more realistic and diverse than simulation data, better reflecting the complexities of the manufacturing process. Through continuous training and optimization, the intelligent controller gradually improves its robustness and adaptability, ensuring excellent performance under various operating conditions.

[0181] Working principle:

[0182] First, a detailed drawing of the double-blade semi-open sewage pump impeller is drawn, and a mold suitable for the shape of the impeller is made. Subsequently, the material for making the impeller is introduced into the mold cavity 14. The first drive motor 41 drives the first telescopic rod 42 to extend, thereby pushing out the bottom fixed plate 46, so that the upper film 12 and the lower film 13 can be accurately matched. A certain amount of time is maintained to allow the material to cool and solidify to ensure that the molded product has sufficient hardness and strength. Subsequently, the first drive motor 41 contracts, and under the action of the first telescopic controller 51, the bottom plate 55 is pushed out. When the rear end of the slide 52 moves to the front end of the slide 54, the first rotating motor 53 drives the fixed rod 57 to rotate, thereby realizing the flipping of the molding plate 56, which facilitates pouring the solidified impeller into the bearing in the heating furnace 61. The metal plate 66 is connected, and then the top cover on the top of the heating furnace 61 is covered. The interior of the heating furnace 61 is vacuumed by external equipment to prevent the impeller surface from being oxidized. Then the second rotating motor 63 drives the rotating rod 65 to rotate, so that the impeller is heated more evenly. Then the impeller is placed in the cooling box 71, and the circulating pump 73 is immediately turned on. The coolant circulates in the condenser 74 to achieve a cooling effect. The cooled impeller is placed between the two clamping plates 87. The second telescopic controller 86 contracts the clamping plates 87 to complete the clamping of the impeller. Then the second telescopic rod 85 is driven out by the second drive motor 84 to facilitate the polishing of the impeller surface by the polishing component 9 to ensure that the impeller surface is flat and smooth. At this point, the entire workflow is completed.

[0183] The above-mentioned front, back, left, right, up and down are all based on the Figure 1 As a benchmark, according to the person's observation perspective, the side of the device facing the observer is defined as the front, the left side of the observer is defined as the left, and so on.

[0184] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the scope of protection of the present invention.

[0185] It should be noted that the device structure and drawings of the present invention mainly describe the principles of the present invention. In terms of the technology of the design principle, the settings of the device's power mechanism, power supply system, and control system are not fully described. However, those skilled in the art can clearly understand the details of its power mechanism, power supply system, and control system on the premise that they understand the principles of the above invention. The control method of the application document is automatic control through a controller, and the control circuit of the controller can be implemented by simple programming by those skilled in the art.

[0186] The standard parts used can be purchased from the market and can be customized according to the description in the specification and drawings. The specific connection methods of each part adopt conventional means such as mature bolts, rivets, welding, etc. in the existing technology. The machinery, parts and equipment all adopt conventional models in the existing technology, and the components known to technical personnel in this field, their structures and principles can be known to these technical personnel through technical manuals or through conventional experimental methods.

[0187] The above describes the embodiments of the present invention in detail in conjunction with the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions and variations of these embodiments without departing from the principles and spirit of the present invention are still within the scope of protection of the present invention.

Claims

1. A manufacturing method for a double-blade semi-open sewage pump impeller manufacturing device, characterized in that: The specific steps of the manufacturing method are as follows: Step 1: First, draw a detailed drawing of the double-blade semi-open sewage pump impeller, make a mold suitable for the shape of the impeller, then introduce the material for manufacturing the impeller into the mold cavity (14), and the first drive motor (41) drives the first telescopic rod (42) to extend, thereby pushing out the bottom fixed plate (46), so that the upper mold (12) and the lower mold (13) can be accurately matched, and keep a certain time for the material to cool and solidify to ensure that the molded product has sufficient hardness and strength. Then, the first drive motor (41) contracts, and under the action of the first telescopic controller (51), the bottom plate (55) is pushed out. When the rear end of the slide plate (52) moves to the front end of the chute (54), the first rotating motor (53) drives the fixed rod (57) to rotate, thereby realizing the flipping of the molding plate (56), making it convenient to pour the solidified impeller into the receiving metal plate (66) in the heating furnace (61); Step 2: When the solidified impeller is poured onto the receiving metal plate (66) in the heating furnace (61), the top cover of the heating furnace (61) is closed, and the interior of the heating furnace (61) is vacuumed by an external device to prevent the impeller surface from being oxidized. Then, the second rotating motor (63) drives the rotating rod (65) to rotate so that the impeller is heated more evenly. Then, the impeller is placed in the cooling box (71), and the circulating pump (73) is immediately turned on. The coolant circulates in the condenser (74) to achieve a cooling effect. Step 3: The cooled impeller is placed between the two clamping plates (87), and the second telescopic controller (86) contracts the clamping plates (87) to complete the clamping of the impeller. Then, the second telescopic rod (85) is pushed out by the second drive motor (84), so that the impeller surface can be polished by the polishing assembly (9) to ensure that the impeller surface is flat and smooth. In step 1, when manufacturing the double-blade semi-open sewage pump impeller, an intelligent algorithm combining the deep deterministic policy gradient algorithm and the dual deep Q network is introduced to optimize the manufacturing process of the double-blade semi-open sewage pump impeller. The specific process is as follows: Step 1: Create an environment model Creating a simulation environment Simulation goal: To simulate all stages of the manufacturing process for a two-blade semi-open sewage pump impeller, including material introduction, cooling and solidification, and forming plate flipping; Deep neural network model structure: Set up input layers, hidden layers, and output layers in the deep neural network model to model and predict various state variables in the manufacturing process; Training a deep neural network model: Data Collection: Historical and experimental data were collected. The dataset included all state variables in each manufacturing process, recording the entire process from material introduction to forming plate flipping. The dataset also included material temperature, cooling time, curing hardness, drive motor position, telescopic rod position, and forming plate angle. The experimental and historical data were used to train the deep neural network model. Step 2: Designing the Intelligent Controller Design of intelligent controller based on dual deep Q network algorithm Strategy network: Receives the current state input of the double-blade semi-open sewage pump impeller manufacturing process and outputs corresponding control actions. These actions are used to adjust various control parameters in the manufacturing process to achieve the best manufacturing effect. Step 3: Train the Smart Controller Training in a simulation environment Simulation run: start running in the simulation environment and set the initial state; The simulation environment simulates all stages of the real manufacturing process, including material introduction, cooling and solidification, and molded plate flipping; Action selection: At each time step, the intelligent controller selects an action a based on the current state s. This action is generated by the policy network, which outputs the control parameters motor speed and telescopic rod position adjustment. Execute action: Execute the selected action a, make the manufacturing equipment operate according to the control parameters, and obtain the next state s` and immediate reward r; the immediate reward is based on the effect of the current action, material introduction uniformity, cooling effect and molding quality.

2. The method according to claim 1, characterized in that The manufacturing device comprises a workbench (1), the front end of the workbench (1) is fixedly connected to a connecting plate (3), the surface of the workbench (1) is fixedly connected to a fixing frame (2), the top of the fixing frame (2) is provided with a pressing mechanism (4), the left and right ends of the workbench (1) are respectively provided with a moving flipping mechanism (5) and a moving clamping mechanism (8), the left and right ends of the connecting plate (3) are respectively provided with a rotating heating treatment mechanism (6) and a cooling mechanism (7), the pressing mechanism (4) comprises a first driving motor (41) fixedly mounted on the top of the fixing frame (2), the output end of the first driving motor (41) is fixedly connected to a first telescopic rod (42), and the first telescopic rod (42) is away from the first driving motor (41). One end of the motor (41) is fixedly connected to a receiving plate (43), and the four corners of the bottom of the receiving plate (43) are fixedly connected to telescopic columns (44). One end of the four telescopic columns (44) away from the receiving plate (43) is fixedly connected to a fixed plate (46). The bottom surface of the fixed plate (46) is fixedly connected to an upper mold (12). The movable flip mechanism (5) includes a first telescopic controller (51). The output end of the first telescopic controller (51) is fixedly connected to a bottom plate (55) via a connecting shaft. A forming plate (56) is provided on the surface of the forming plate (56). A cavity (14) is opened on the surface of the forming plate (56). The bottom surface of the cavity (14) is fixedly mounted with a lower mold (13). The circumferential surfaces of the four telescopic columns (44) are sleeved with telescopic springs (45), and the upper and lower ends of the telescopic springs (45) are fixedly mounted on the bottom surface of the receiving plate (43) and the surface of the fixed plate (46), respectively. The upper mold (12) and the lower mold (13) are on the same vertical line.

3. The method according to claim 2, wherein: A through slot (10) is provided at the right end of the surface of the workbench (1), a grinding assembly (9) is fixedly mounted on the rear side of the right end of the workbench (1) via a motor seat, and a rear plate (11) is fixedly mounted on the rear bottom end of the fixing frame (2).

4. The method according to claim 2, wherein: Side plates (58) are fixedly installed on the surface of the bottom plate (55) and located on the left and right sides of the forming plate (56). The outer front end of the left side plate (58) is fixedly connected to the first rotating motor (53). The output end of the first rotating motor (53) is fixedly installed with a fixing rod (57). The fixing rod (57) is fixedly installed on the inner front end of the forming plate (56). A bearing seat adapted to the fixing rod (57) is fixedly installed on the surface of the right side plate (58). The bottom surface of the bottom plate (55) is fixedly connected to the slide plate (52). The surface of the workbench (1) is provided with a slide groove (54) adapted to the slide plate (52). The first telescopic controller (51) is fixedly installed on the outer surface of the rear plate (11).

5. The method according to claim 2, wherein: The rotary heating treatment mechanism (6) includes a heating furnace (61), a top cover is provided on the top of the heating furnace (61), and a vacuum exhaust hole (62) is provided on one end of the surface of the top cover. The heating furnace (61) is fixedly installed at the inner left end of the connecting plate (3), a heating coil (67) is fixedly installed on the inner wall of the heating furnace (61), a temperature display controller (64) is fixedly installed on the front outer surface of the heating furnace (61), a second rotating motor (63) is fixedly connected to the bottom outer surface of the heating furnace (61), the output end of the second rotating motor (63) is fixedly connected to a rotating rod (65), and the end of the rotating rod (65) away from the second rotating motor (63) is fixedly connected to a receiving metal plate (66).

6. The method according to claim 2, wherein: The cooling mechanism (7) includes a cooling box (71) fixedly mounted inside the right end of the connecting plate (3), the right side surface of the cooling box (71) is fixedly connected to a box body (72), a circulating pump (73) is provided on the top of the box body (72), an output end of the circulating pump (73) is fixedly mounted with a water inlet pipe (75), one end of the water inlet pipe (75) is arranged inside the box body (72), the other end of the water inlet pipe (75) is fixedly connected to a condenser (74), one end of the condenser (74) away from the water inlet pipe (75) is fixedly connected to a return pipe (76), one end of the return pipe (76) away from the condenser (74) is arranged in the box body (72), the condenser (74) is fixedly mounted on the inner wall of the cooling box (71), and the condenser (74) is arranged in a spiral downward shape.

7. The method according to claim 2, wherein: The movable clamping mechanism (8) comprises a limiting groove (81) provided at the right end of the surface of the workbench (1), wherein the limiting grooves (81) are provided with two, a plurality of sliders are slidably connected in the two limiting grooves (81), and two mounting plates (83) are fixedly connected on the surfaces of the plurality of sliders, a second drive motor (84) is fixedly installed between the two mounting plates (83) via a drive motor seat, a telescopic push rod (82) is provided inside the limiting groove (81), one end of the telescopic push rod (82) is fixedly connected to the slider, the output end of the second drive motor (84) is fixedly connected to a second telescopic rod (85), the end of the second telescopic rod (85) away from the second drive motor (84) is fixedly connected to a second telescopic controller (86), the upper and lower ends of the second telescopic controller (86) are both provided with clamping plates (87), and the surface of the clamping plate (87) is provided with an anti-slip pad.

8. The manufacturing method of the double-blade semi-open sewage pump impeller manufacturing device according to claim 1, characterized in that: Step 1 also includes: Simulation tools: Use Python and the SimPy library to create a simulation environment, simulate the entire manufacturing process, and collect data on different state variables. SimPy is a Python library for event-driven simulation, suitable for creating discrete event simulation models of manufacturing processes. Simulation steps: Material import: simulate the process of injecting material into the mold; Cooling and solidification: simulate the cooling and solidification process of materials; Forming plate flipping: simulates the process of flipping the forming plate to release the curing impeller; State variables Material temperature T m : Indicates the current temperature of the material, ranging from room temperature to curing temperature; Cooling time t cool : Indicates the material cooling time, set the preset range Curing hardness H cure : Indicates the curing hardness of the material, ranging from 0 uncured to 1 fully cured; Drive motor position P motor : Indicates the position of the drive motor, ranging from 0 starting position to 1 end position; Telescopic rod position P rod : Indicates the position of the telescopic rod, ranging from 0 for contraction to 1 for full extension; Forming plate angle A plate : Indicates the angle of the forming plate, ranging from 0 degrees horizontal to 90 degrees vertical; Using deep neural networks to represent environmental dynamics to monitor and control state variables in the impeller manufacturing process; Input layer: including all state variables, material temperature T m , cooling time t cool , cured hardness H cure , driving motor position P motor , telescopic rod position P rod , forming plate angle A plate ; Hidden layer: two hidden layers, each containing 64 neurons, with ReLU activation function; Output layer: predicts all state variables at the next moment; optimizes control parameters at each stage of the manufacturing process to improve manufacturing efficiency and quality. Use optimizer and loss function: Use the Adam optimizer and mean square error (MSE) loss function to optimize the deep neural network model to ensure that the model can accurately predict and simulate state changes in the manufacturing process; Adam optimizer formula: m t =β1m t-1 +(1-β1)g t m t : The first-order moment estimate at time t, momentum; this is the momentum at the current moment, combined with the momentum m at the previous moment t-1 and the current gradient g t ; v t : The second-order moment estimate at time t, the square of the gradient; this is the square of the gradient at the current moment, combined with the second-order moment estimate v at the previous moment t-1 and the square of the current gradient Momentum m t Bias-corrected estimate of the momentum m t The estimation was performed after bias correction, taking into account the exponential decay of β1; Second moment estimate v t The bias-corrected estimate of the second-order moment estimate v t The bias-corrected estimates take into account the exponential decay of β2; θ t : The parameter value at time t; is the updated parameter value, combining the correction estimate of momentum and second-order moment, using the learning rate α and a small constant ε to prevent division by zero errors; g t : Gradient at time t This is the gradient value at the current moment; β1: exponential decay rate of the first-order moment estimate; β2: exponential decay rate of the second-order moment estimate; α: learning rate; this is the coefficient used to adjust the parameter update step size; ε: A constant to prevent division by zero errors; usually set to a very small value The loss function is the mean square error (MSE), which is used to minimize the error between the predicted value and the true value to ensure the accuracy of the model prediction; yi: true value; Predicted value; n: number of samples; The dual-depth Q-network algorithm combines a policy network and a value network to determine the optimal action and evaluate the value of the action, respectively. An intelligent controller is designed using the dual-depth Q-network algorithm to optimize the control parameters of the impeller manufacturing process for a two-blade semi-open sewage pump. Step 2 also includes: Input: Current state s, including the following state variables: Material temperature T m , cooling time t cool , cured hardness H cure , driving motor position P motor , telescopic rod position P rod , forming plate angle A plate ; structure: Input layer: 6 state variables; Hidden layer: two hidden layers, 64 neurons in each layer, and the activation function is ReLU; Output layer: controls the action, the activation function is tanh; Value network: Evaluates the value of executing a specific action in the current state of the double-blade semi-open sewage pump impeller manufacturing process, thereby guiding the policy network to select the optimal action in the double-blade semi-open sewage pump impeller manufacturing process; Input state s: the same 6 state variables as the policy network input material temperature T m , cooling time t cool , cured hardness H cure , driving motor position P motor , telescopic rod position P rod , forming plate angle A plate ; Output action a: The control action output by the strategy network, which is the motor speed and telescopic rod position adjustment during the manufacturing process of the double-blade semi-open sewage pump impeller; Output: Q value, which represents the value of a given state and action in the manufacturing process of a double-blade semi-open sewage pump impeller. The Q value is used to evaluate the value of the current action to help the policy network optimize its output. structure: Input layer: a combination of state and action; Hidden layer: two hidden layers, 64 neurons in each layer, and the activation function is ReLU; Output layer: Q value; Optimizing Q-value estimation using a dual deep Q-network algorithm The dual-deep Q network reduces the bias in Q-value estimation by using two independent networks to calculate action selection and action value respectively. During the impeller manufacturing process, immediate rewards can be set based on molding quality, cooling speed, and energy consumption indicators. When the material is introduced evenly, the cooling time is appropriate, and the solidification hardness meets the requirements, high rewards are given; otherwise, low rewards or penalties are given. The discount factor in the calculation is used to weigh the importance of immediate rewards and future rewards. In the manufacturing process, it is necessary to weigh the current control action and the future molding quality. Target Q network: Use an independent target Q network to calculate the target Q value to reduce estimation bias; the target Q value of different state and action combinations is calculated by the target Q network, thereby guiding the policy network to optimize the control action; The state s` and action a` of the next moment are predicted during the manufacturing process through the current control action and state change. The material temperature and cooling hardness of the next moment are predicted through the current material temperature and cooling time. The drive motor position and telescopic rod position of the next moment are predicted through the current drive motor position and telescopic rod position. The target value calculation formula of the dual deep Q network is: y=rγQ`(s`,argmax a `Q(s`,a`|θ Q )|θ Q` ) in: y: target value, which is the Q-value update target and represents the value estimation target of the current state-action pair; r: immediate reward, the reward obtained after performing the action in the current state; γ: Discount factor used to discount the importance of future rewards, ranging from 0 to 1, Q`: Target Q network, an independent Q network used to calculate the target value to reduce estimation bias; s`: The state at the next moment, the state to be transferred to after executing the current action; argmax a `Q(s`,a`|θ Q )|θ Q` ): The strategy for selecting an action is to select the action a` that maximizes the Q value in the next state s`; where Q(s`,a`|θ Q )|θ Q` ) represents the Q value calculated using the current Q network; θ Q : Parameters of the current Q network, parameters of the current Q network, used to calculate the Q value; θ Q` : Parameters of the target Q network, used to calculate the target Q value; Step 3 also includes: Initialize the parameters of the Deep Deterministic Policy Gradient and Dual Deep Q-Network algorithms Parameter initialization: Before training begins, the weights of the policy network and the value network are randomly initialized to break the symmetry and make the network learn different features; The target network and Q' copy the initial weights of the policy network and value network: the target network is used to stabilize the training process. Its initial weights are directly copied from the policy network and value network. As training progresses, the weights of the target network will be slowly updated; Set the learning rate α and discount factor γ: Learning rate α: controls the step size of parameter updates to ensure stable and gradual learning; Discount factor γ: weighs the importance of immediate rewards and future rewards; Experience replay: Use the experience replay mechanism to store state transitions (s, a, r, s`); the experience replay buffer stores these state transitions so that small batches of data can be extracted from them for network updates later; this mechanism can break data correlation and enhance training effects; Network Update Value network loss function: Calculate the loss function of the value network to optimize the parameters of the value network; Loss function: L(θ Q )=E s,a,r,s `[(Q(s,a|θ Q )-y) 2 ] L(θ Q ): The loss function of the value network, which is used to measure the error between the predicted value of the value network Q and the target value y; θ Q : Parameters of the value network, used to represent the weights and biases of the value network. These parameters need to be optimized through training; E s,a,r,s `: Expectation, which represents the expected value of all states, actions, rewards and next states; Q(s,a|θ Q ): The output of the value network, the estimated value of the value network Q under a given state s and action a, which represents the expected return in state s after taking action a; y: target value, the target value to be learned by the value network, used to update the parameters of the value network; Policy network gradient: Calculate the gradient of the policy network to optimize the parameters of the policy network; Gradient formula: Policy network gradient, which represents the objective function J versus the policy network parameter θ μ The gradient of is used to update the parameters of the policy network; E s : Expectation, which represents the expected value of all states s; The gradient of the value network to the action represents the gradient of the value network Q to the action a given state s and action a; a = μ(s|θ): the action generated by the policy network, which represents the action a generated by the policy network in a given state s; The gradient operator represents the differentiation or derivation of a function to calculate its gradient; θ μ : The parameters of the policy network represent the weights and bias parameters of the policy network μ, which need to be optimized through training; μ: Policy network represents the function of the policy network, input state s, and output control action a; s: state, which represents the state of the current environment and serves as the input of the policy network; μ(s|θ μ ): The output of the policy network, indicating that in a given state s, the policy network μ is based on the parameter θ μ Generated control action a; Step 4: Real-time Control and Adjustment Use trained intelligent controller Ensure that the intelligent controller model has been fully trained and verified in the simulation environment. This includes continuous training and testing to ensure that the intelligent controller can stably generate reasonable control actions in the simulation environment and perform well in various simulation scenarios. Deploy the trained intelligent controller model to the actual manufacturing equipment control system, connect the intelligent controller model to the hardware interface of the actual equipment, and integrate it into the existing control system to enable it to control the drive motor, telescopic rod, and forming plate components. The intelligent controller generates control actions based on real-time monitored state variables such as material temperature, telescopic rod position, and forming plate angle. These control actions act directly on the equipment to ensure that the equipment operates according to the optimal strategy, thereby ensuring the efficiency and accuracy of the manufacturing process. Sensors installed at various locations monitor the system status in real time. The real-time data collected by the sensors is transmitted to the control system via the data bus. The intelligent controller analyzes this data in real time to determine whether the manufacturing process parameters need to be adjusted. The intelligent controller selects the optimal control action based on the real-time status, calculates the optimal motor speed, telescopic rod position, and forming plate angle, and executes these actions immediately.

Images