A water pump impeller and a manufacturing device and method thereof

Abstract

The invention discloses a water pump impeller and a manufacturing device and method thereof. The invention relates to the technical field of water pump impeller manufacturing, comprising a machine base and a cooling mechanism arranged at a position near the left side of the upper surface thereof, a debris air sweeping mechanism is arranged at a position near the upper right of the upper surface of the machine base, a collecting mechanism is arranged on the front of the debris air sweeping mechanism, the cooling mechanism comprises a cooling box, the cooling box is movably connected with a placing rack through a cooling groove, a cooling channel is opened inside the cooling box, the advantage of the invention is that: through the cooling mechanism, a circulating pump draws water into the cooling channel, and then impurities in the water are filtered through a filter element of the filter box, the filtered water enters the interior of the refrigeration box for cooling, and finally the cooled water flows back from the water outlet pipe to the interior of the cooling channel to cool the air in the cooling groove, thereby cooling the casted water pump impeller, accelerating the cooling speed, and improving the manufacturing efficiency of the water pump impeller.

Description

Technical Field

[0001] The invention relates to the technical field of water pump impeller manufacturing, and in particular to a water pump impeller and a manufacturing device and method thereof. Background Art

[0002] The impeller of the water pump is made of cast iron. The blades on the impeller of the water pump play a major role. The shape and size of the impeller of the water pump are closely related to the performance of the water pump. The impeller of the water pump can generally be divided into two types: single suction and double suction. The single suction impeller absorbs water on one side, and the impeller of the small flow water pump is mostly of this style. The double suction impeller absorbs water on both sides, and the impeller of the large flow water pump adopts the double suction impeller. When the impeller of the water pump is manufactured in large quantities, it is generally cast by mold;

[0003] After searching, the applicant found that a Chinese patent disclosed "a water pump impeller manufacturing device and method", and its publication (announcement) number is "CN114523106A". This patent mainly uses the negative pressure to extract the air at the suction hole, so that the push plate can fix the water pump impeller, and then by rotating the push plate, the push plate drives the water pump impeller to rotate, so that the water pump impeller can shake off the molding sand attached to the surface, which is convenient for subsequent fine processing and does not require secondary cleaning. However, this patent cannot artificially cool the water pump impeller that has just been cast, and waiting for it to cool naturally is time-consuming. For this reason, we propose a water pump impeller and a manufacturing device and method thereof. Summary of the invention

[0004] The object of the present invention is to provide a water 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 water pump impeller manufacturing device, comprising a machine base and a cooling mechanism arranged at a position near the left side of the upper surface thereof, a debris air sweeping mechanism is arranged at a position near the upper right of the upper surface of the machine base, a collecting mechanism is arranged on the front of the debris air sweeping mechanism, the cooling mechanism comprises a cooling box, the cooling box is movably connected to a placement rack through a cooling groove, a cooling channel is opened inside the cooling box, one end of the cooling channel is fixedly connected to a circulation pump through a return pipe, the circulation pump is fixedly connected to a refrigeration box through a filter box, and the other end of the cooling channel is connected to a water outlet pipe The debris air sweeping mechanism is fixedly connected to the refrigeration box, and includes a motor and a transparent sealed box. A turntable is fixedly installed on the output shaft of the motor. A slot is provided on the front of the transparent sealed box. A cover plate is movably installed on the upper end of the transparent sealed box. An air compressor is fixedly installed on the upper surface of the cover plate near the left side. The air compressor is fixedly connected to a filter through an air storage tank. The filter is fixedly connected to an air pipeline through a pressure regulating valve. A nozzle is fixedly installed on the bottom of the air pipeline. The collecting mechanism includes an air suction pump, and the air suction pump is fixedly connected to an adapter cover through a dust suction pipe and a docking pipe. An interception filter is inserted into the interior of the adapter cover.

[0006] As a further solution of the present invention: a mounting groove is provided inside the base near the upper right position, and a plug-in groove is provided outside the mounting groove.

[0007] As a further solution of the present invention: the bottom of the cooling box is fixedly installed on the upper surface of the base near the upper left position, a sealing cover is movably installed on the upper end of the cooling box, the cooling groove is opened inside the cooling box, the placement rack is plugged into the inside of the cooling groove, air holes are opened on the outer wall of the placement rack, an impeller placement plate is arranged inside the placement rack, and a water inlet pipe is fixedly installed on the left side of the cooling box, and the water inlet pipe is communicated with the cooling channel.

[0008] As a further solution of the present invention: one end of the return water pipe is fixedly installed at one end of the cooling channel, the other end of the return water pipe is fixedly connected to the return water end of the circulating pump, one end of the filter box is fixedly connected to the water outlet end of the circulating pump, a filter element is inserted into the filter box, one end of the refrigeration box is fixedly connected to the other end of the filter box, the other end of the refrigeration box is fixedly connected to one end of the outlet pipe, and the other end of the outlet pipe is fixedly installed at the other end of the cooling channel.

[0009] As a further solution of the present invention: the bottom of the motor is fixedly installed inside the installation slot, and a plug-in component is fixedly installed on the bottom of the transparent sealing box, and the plug-in component is plugged into the inside of the plug-in slot.

[0010] As a further solution of the present invention: the air storage tank is fixedly connected to the air outlet pipe of the air compressor, the filter is fixedly connected to the air outlet pipe of the air storage tank, the pressure regulating valve is fixedly installed on the air outlet pipe of the filter, the upper end of the air supply pipeline is fixedly connected to the pressure regulating valve, and the nozzle is located inside the transparent sealing box near the upper end.

[0011] As a further solution of the present invention: the collection mechanism includes an air suction pump, which is fixedly installed on the upper surface of the base near the lower right position, one end of the dust suction pipe is fixedly installed on the suction end of the air suction pump, one end of the docking pipe is fixedly connected to the other end of the dust suction pipe, the other end of the docking pipe is fixedly connected to one end of the adapter cover, and the other end of the adapter cover is fixedly connected to the slot.

[0012] In addition, the present invention also provides a method for manufacturing a water pump impeller of a water pump impeller manufacturing device, comprising the following steps:

[0013] Step 1: First, insert the placement rack into the cooling tank, place the newly cast water pump impeller in the placement tank of the impeller placement plate, then introduce cold water from the water inlet pipe into the cooling channel, the cold water is pumped into the cooling channel from the return pipe by the circulating pump, and then enters the filter box from the water outlet of the circulating pump. The impurities and dirt in the water are filtered by the filter element, and the filtered water enters the refrigeration box to be refrigerated, and then flows back from the water outlet pipe to the inside of the cooling channel. The air in the circulating cooling tank of the circulating pump is refrigerated, and the newly cast water pump impeller in the placement rack is cooled;

[0014] Step 2: Open the cover plate, place the cooled water pump impeller on the turntable, then cover the cover plate on the upper end of the transparent sealing box, start the motor, make the water pump impeller move in a circle along the direction of the turntable, compress the air through the air compressor, and then store it in the air storage tank, then filter the impurities and moisture in the air through the filter to ensure the cleanliness of the purge air, then adjust the pressure of the compressed air through the pressure regulating valve to meet the purge requirements, and finally spray gas through the nozzle at the lower end of the gas pipeline to purge the debris on the surface of the water pump impeller;

[0015] Step 3: When blowing away the debris on the surface of the water pump impeller, turn on the suction pump, the debris is sucked in from the adapter cover, the debris is adsorbed by the interception filter, and the gas and fine dust enter the internal air path of the suction pump through the docking tube and are discharged.

[0016] In addition, the present invention also provides a water pump impeller of a water pump impeller manufacturing device, wherein an impeller mechanism is arranged inside the cooling mechanism, and the impeller mechanism includes a rim, and the rim is arranged in a placement groove of an impeller placement plate, and a hub is fixedly installed at the middle position of the upper surface of the rim, and blades are fixedly installed on the outer side of the hub, and rust-proof blades are fixedly installed on the outer wall of the blade.

[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 adopts a cooling mechanism to draw water into the cooling channel through a circulating pump, and then the impurities in the water are filtered through the filter element of the filter box. The filtered water enters the interior of the refrigeration box for cooling, and finally the cooled water flows back from the outlet pipe to the interior of the cooling channel to cool the air in the cooling tank, thereby cooling the cast water pump impeller, accelerating the cooling speed, and improving the manufacturing efficiency of the water pump impeller;

[0019] 2. The present invention provides a debris air sweeping mechanism. The motor rotates to drive the water pump impeller on the turntable to rotate. The air is compressed by the air compressor and then stored in the air storage tank. The impurities and moisture in the air are then filtered through the filter to ensure the cleanliness of the air for sweeping. The pressure of the compressed air is then adjusted by the pressure regulating valve to meet the sweeping requirements. Finally, the gas is sprayed through the nozzle at the lower end of the gas pipeline to sweep the debris on the surface of the water pump impeller. The water pump impeller can be swept in all directions without dead angles.

[0020] 3. The present invention provides a collection mechanism, and the suction pump sucks the debris in the transparent sealing box from the adapter cover, and the debris is adsorbed by the interception filter. The gas and fine dust enter the internal air path of the suction pump through the docking tube and are then discharged. The debris air sweeping mechanism can cooperate to thoroughly clean the debris on the water pump impeller.

[0021] 4. The present invention uses an impeller mechanism to cut impurities such as fibers in the working environment of the water pump impeller through the rust-proof blades on the blades, thereby preventing the fibers from being entangled in the water pump impeller and affecting the operation of the water pump impeller.

[0022] 5. Precise cooling control: By installing temperature sensors and flow sensors at multiple locations in the cooling tank, cooling channel and water pump impeller, the temperature and flow data at each location are monitored in real time. These data are transmitted to the central control module in real time through the wireless communication module to ensure that the system can respond to temperature changes in a timely manner and achieve precise cooling. Using deep reinforcement learning combined with model predictive control algorithms, future temperature changes are predicted based on real-time and historical data, and the cooling water flow and circulating pump speed are optimized to ensure efficient and uniform cooling.

[0023] 6. Improve cooling efficiency: Through deep reinforcement learning algorithms, historical data and real-time data are learned to optimize the control strategy, so that the system can adaptively adjust the cooling water flow and circulation pump speed to maximize cooling efficiency. Model predictive control (MPC) uses deep reinforcement learning to optimize the control strategy for real-time prediction and correction to form a closed-loop control system to ensure the accuracy of the system in practical applications.

[0024] 7. Energy saving and consumption reduction: In the optimization objective function, the temperature error and control input energy consumption are balanced through the state weight matrix (Q) and the control input weight matrix (R), ensuring that the system can achieve efficient cooling while reducing energy consumption as much as possible. Based on the prediction and optimization results of deep reinforcement learning combined with the model predictive control algorithm, the flow rate of cooling water, the operating speed of the circulation pump and the refrigeration effect of the refrigeration box are automatically adjusted to further optimize energy consumption and ensure the energy saving effect of the system.

[0025] 8. System stability and reliability: Noise filtering and data smoothing: Smooth the collected temperature and flow data to eliminate noise and outliers, ensure data accuracy and consistency, and improve system stability. Model predictive control performs short-term predictions and corrections based on real-time data, adjusts control strategies in a timely manner, ensures the stability and uniformity of the cooling process, and reduces system fluctuations.

[0026] 9. Improved intelligence level: Through the central control module integrating sensor data and deep reinforcement learning combined with the analysis results of the model predictive control algorithm, the system can realize a fully automated intelligent cooling process, reduce manual intervention, and improve production efficiency. The deep reinforcement learning algorithm is constantly learning and improving, so that the control strategy of the system is gradually optimized, which can adapt to different working conditions and improve the intelligence level of the system.

[0027] 10. Significantly improved cooling effect: By automatically adjusting the cooling water flow, circulating pump speed and refrigeration effect of the refrigeration box, it ensures that each link in the cooling process can reach the best state, so that the cooling effect is significantly improved. Uniform temperature: Through intelligent adjustment, the system can maintain the uniformity of the temperature in the cooling tank, avoid overcooling or overheating, and ensure the cooling effect of the impeller. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the overall three-dimensional structure in an embodiment of the present invention;

[0029] Figure 2 It is a schematic diagram of some structures in an embodiment of the present invention;

[0030] Figure 3 An exploded view of a portion of the cooling mechanism in an embodiment of the present invention;

[0031] Figure 4 A partial structural schematic diagram of a cooling mechanism in an embodiment of the present invention;

[0032] Figure 5 Schematic diagram of the three-dimensional structure of the impeller mechanism in the embodiment of the present invention;

[0033] Figure 6 This is one of the partial structural schematic diagrams of the debris air sweeping mechanism in the embodiment of the present invention;

[0034] Figure 7 This is a second partial structural diagram of the debris air sweeping mechanism in an embodiment of the present invention;

[0035] Figure 8 The third schematic diagram of the partial structure of the debris air sweeping mechanism in the embodiment of the present invention;

[0036] Fig. 9 It is a schematic diagram of the three-dimensional structure of the collecting mechanism in an embodiment of the present invention.

[0037] Fig.10 Add intelligent algorithm flow chart to the present invention

[0038] In the figure: 1, base; 2, mounting slot; 3, plug-in slot; 4, cooling mechanism; 401, cooling box; 402, water inlet pipe; 403, cooling slot; 404, placement rack; 405, air vent; 406, impeller placement plate; 407, sealing cover; 408, cooling channel; 409, return pipe; 410, circulation pump; 411, filter box; 412, filter element; 413, refrigeration box; 414, water outlet pipe; 5, debris air sweeping mechanism; 501, motor; 502, Turntable; 503, connector; 504, transparent sealing box; 505, slotting; 506, cover plate; 507, air compressor; 508, air storage tank; 509, filter; 510, pressure regulating valve; 511, gas pipeline; 512, nozzle; 6, collecting mechanism; 61, suction pump; 62, dust suction pipe; 63, docking pipe; 64, adapter cover; 65, interception filter; 7, impeller mechanism; 71, rim; 72, hub; 73, blade; 74, rust-proof blade. DETAILED DESCRIPTION

[0039] 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.

[0040] 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.

[0041] Please refer to the attached Figure 1 -Attached Fig. 9The present invention provides a water pump impeller manufacturing device, comprising a machine base 1 and a cooling mechanism 4 arranged on the upper surface thereof near the left side, a debris air sweeping mechanism 5 is arranged on the upper surface of the machine base 1 near the upper right position, a collecting mechanism 6 is arranged on the front of the debris air sweeping mechanism 5, the cooling mechanism 4 comprises a cooling box 401, the cooling box 401 is movably connected to a placement rack 404 through a cooling groove 403, a cooling channel 408 is opened inside the cooling box 401, one end of the cooling channel 408 is fixedly connected to a circulation pump 410 through a return pipe 409, the circulation pump 410 is fixedly connected to a refrigeration box 413 through a filter box 411, and the other end of the cooling channel 408 is fixedly connected to the refrigeration box 413 through a water outlet pipe 414, the debris air sweeping mechanism 5 comprises A motor 501 and a transparent sealed box 504, wherein a turntable 502 is fixedly mounted on the output shaft of the motor 501, a slot 505 is provided on the front of the transparent sealed box 504, a cover plate 506 is movably mounted on the upper end of the transparent sealed box 504, an air compressor 507 is fixedly mounted on the upper surface of the cover plate 506 near the left side, the air compressor 507 is fixedly connected to a filter 509 via an air storage tank 508, the filter 509 is fixedly connected to an air pipeline 511 via a pressure regulating valve 510, a nozzle 512 is fixedly mounted on the bottom of the air pipeline 511, the collecting mechanism 6 comprises an air suction pump 61, the air suction pump 61 is fixedly connected to an adapter cover 64 via a dust suction pipe 62 and a docking pipe 63, an interception filter 65 is inserted into the interior of the adapter cover 64.

[0042] In one embodiment of the present invention, a mounting groove 2 is provided in the base 1 near the upper right position, and a plug-in groove 3 is provided on the outer side of the mounting groove 2 .

[0043] In one embodiment of the present invention: the bottom of the cooling box 401 is fixedly installed on the upper surface of the base 1 near the upper left position, a sealing cover 407 is movably installed on the upper end of the cooling box 401, a cooling groove 403 is opened inside the cooling box 401, a placement rack 404 is plugged into the cooling groove 403, an air vent 405 is opened on the outer wall of the placement rack 404, an impeller placement plate 406 is arranged inside the placement rack 404, and a water inlet pipe 402 is fixedly installed on the left side of the cooling box 401, and the water inlet pipe 402 is communicated with the cooling channel 408.

[0044] In one embodiment of the present invention: one end of the return pipe 409 is fixedly installed at one end of the cooling channel 408, the other end of the return pipe 409 is fixedly connected to the return end of the circulation pump 410, one end of the filter box 411 is fixedly connected to the water outlet end of the circulation pump 410, a filter element 412 is inserted into the filter box 411, one end of the refrigeration box 413 is fixedly connected to the other end of the filter box 411, the other end of the refrigeration box 413 is fixedly connected to one end of the outlet pipe 414, and the other end of the outlet pipe 414 is fixedly installed at the other end of the cooling channel 408.

[0045] In one embodiment of the present invention: the bottom of the motor 501 is fixedly installed inside the installation slot 2 , and the bottom of the transparent sealing box 504 is fixedly installed with a plug-in component 503 , which is plugged into the plug-in slot 3 .

[0046] In one embodiment of the present invention: the air tank 508 is fixedly connected to the air outlet pipe of the air compressor 507, the filter 509 is fixedly connected to the air outlet pipe of the air tank 508, the pressure regulating valve 510 is fixedly installed on the air outlet pipe of the filter 509, the upper end of the air supply pipeline 511 is fixedly connected to the pressure regulating valve 510, and the nozzle 512 is located inside the transparent sealing box 504 near the upper end.

[0047] In one embodiment of the present invention: the collection mechanism 6 includes an air suction pump 61, which is fixedly installed on the upper surface of the base 1 near the lower right position, one end of the dust suction pipe 62 is fixedly installed on the suction end of the air suction pump 61, one end of the docking pipe 63 is fixedly connected to the other end of the dust suction pipe 62, the other end of the docking pipe 63 is fixedly connected to one end of the adapter cover 64, and the other end of the adapter cover 64 is fixedly connected to the groove 505.

[0048] Example 1, please refer to the attached Figure 1 -Attached Figure 4 The cooling channel 408 is opened on the inner wall of the cooling box 401, so that the cold water after being cooled by the refrigeration box 413 returns to the cooling channel 408 to cool the inner wall of the cooling channel 408, and then cools the air inside the cooling groove 403, so as to achieve the purpose of cooling the water pump impeller after casting. At the same time, insulation plates are fixedly installed on the outer walls of the cooling box 401 and the sealing cover 407 to prevent the loss of internal cold air.

[0049] Example 2, please refer to the attached Figure 1 -Attached Figure 8 The nozzles 512 are distributed along the circumference of the turntable 502, so that when the water pump impeller on the turntable 502 rotates, the nozzles 512 can spray and blow every part of the surface of the water pump impeller to avoid residual debris. At the same time, the filter 509 is set to prevent dust in the external air from being brought into the transparent sealing box 504.

[0050] The present invention also provides a method for manufacturing a water pump impeller of a water pump impeller manufacturing device, comprising the following steps:

[0051] Step 1, first insert the placement rack 404 into the cooling tank 403, place the newly cast water pump impeller in the placement tank of the impeller placement plate 406, then introduce cold water from the water inlet pipe 402 into the cooling channel 408, the cold water is pumped into the cooling channel 408 from the return pipe 409 by the circulation pump 410, and then enters the filter box 411 from the water outlet end of the circulation pump 410, the impurities and dirt in the water are filtered by the filter element 412, the filtered water enters the refrigeration box 413 to be refrigerated, and then flows back to the cooling channel 408 from the water outlet pipe 414, the air in the circulating cooling tank 403 of the circulation pump 410 is refrigerated, and the newly cast water pump impeller in the placement rack 404 is cooled;

[0052] Step 2: Open the cover plate 506, place the cooled water pump impeller on the turntable 502, then cover the cover plate 506 on the upper end of the transparent sealing box 504, start the motor 501, so that the water pump impeller moves in a circle along the direction of the turntable 502, compress the air through the air compressor 507, and then store it in the air storage tank 508, and then filter the impurities and moisture in the air through the filter 509 to ensure the cleanliness of the purge air, and then adjust the pressure of the compressed air through the pressure regulating valve 510 to meet the purge demand, and finally spray gas through the nozzle 512 at the lower end of the gas pipeline 511 to purge the debris on the surface of the water pump impeller;

[0053] Step 3: When cleaning the debris on the surface of the water pump impeller, turn on the suction pump 61, the debris is sucked from the adapter cover 64, the debris is adsorbed by the interception filter 65, and the gas and fine dust enter the internal air path of the suction pump 61 through the docking tube 63 and are discharged.

[0054] Example 3, please refer to the attached Figure 1 -Attached Fig. 9 The opening bottom of the slot 505 is parallel to the surface of the turntable 502. Under the action of the suction pump 61, all the debris inside the transparent sealing box 504 can be drawn out from the adapter cover 64. The interception filter 65 is inserted into the inside of the adapter cover 64. After each operation is completed, the interception filter 65 can be conveniently pulled out from the adapter cover 64 for cleaning or replacement.

[0055] In step 1, when cooling the water pump impeller, deep reinforcement learning is combined with the intelligent algorithm of the model predictive control algorithm to optimize the cooling process of the water pump impeller to improve cooling efficiency, enhance energy saving and consumption reduction, improve the level of intelligence, improve control accuracy and efficiency, and achieve precise cooling control. The specific process is as follows:

[0056] Step 1: Sensor layout and data collection

[0057] Sensor layout

[0058] Temperature sensor: Temperature sensors are installed at multiple locations in the cooling tank, cooling channel and water pump impeller to monitor the temperature data at each location in real time.

[0059] Cooling tank (position A): Install a temperature sensor at the top and bottom of the cooling tank, marked as T A1 and T A2 .

[0060] Top temperature sensor (T A1 ): Installed on the top of the cooling tank to monitor the temperature of the air in the cooling tank.

[0061] Bottom temperature sensor (T A2 ): Installed at the bottom of the cooling tank to monitor the temperature of the water in the cooling tank.

[0062] Cooling channel (position B): Install a temperature sensor at the inlet and outlet of the cooling channel, marked as T B1 and T B2 .

[0063] Inlet temperature sensor (TB1): Installed at the inlet of the cooling channel to monitor the temperature of the cooling water when it enters the cooling channel.

[0064] Outlet temperature sensor (T B2 ): Installed at the outlet of the cooling channel to monitor the temperature of the cooling water when it leaves the cooling channel.

[0065] Impeller (position C): Install a temperature sensor at the center and edge of the pump impeller, marked as T C1 and T C2 .

[0066] Core temperature sensor (T C1 ): Installed in the center of the pump impeller to monitor the temperature of the impeller center.

[0067] Edge temperature sensor (T C2 ): Installed on the edge of the water pump impeller to monitor the temperature of the impeller edge.

[0068] Flow sensor: Install flow sensors on the inlet pipe and return pipe to monitor the flow of cooling water.

[0069] Inlet pipe (position D): Install a flow sensor at the inlet of the inlet pipe, marked as T D . Used to monitor the cooling water flow entering the cooling channel.

[0070] Return pipe (position E): Install a flow sensor at the outlet of the return pipe, marked as T E . Used to monitor the return cooling water flow rate.

[0071] Data acquisition and processing: The data acquisition module collects data from all sensors to the central control module for real-time data acquisition and processing.

[0072] Real-time data transmission

[0073] The data collected by each temperature sensor and flow sensor is transmitted to the central control module in real time through the wireless communication module. The data transmission frequency is set to sample once per second to ensure real-time monitoring and control.

[0074] The data acquisition module performs preliminary processing on the collected data, including data verification and format conversion, to ensure the accuracy and consistency of the data.

[0075] Data storage

[0076] The data after preliminary processing is stored in the database of the central control module. The database adopts an efficient data structure to support fast query and large data storage.

[0077] Temperature data format: (timestamp, sensor location, temperature value)

[0078] Traffic data format: (timestamp, sensor location, flow value)

[0079] Data processing and analysis

[0080] The central control module regularly analyzes the collected data and generates time series data of temperature and flow for model prediction and control strategy optimization.

[0081] Smooth the temperature data to remove noise and outliers. Use a moving average filter:

[0082]

[0083] in, is the smoothed temperature value, T(ti) is the temperature value at sampling time ti, and N is the window size of the moving average.

[0084] Perform statistical analysis on the flow data and calculate the average flow and flow change trend. Calculate the average flow of the inlet pipe and return pipe:

[0085]

[0086] in, and are the average flow rates of the inlet pipe and the return pipe, respectively, and F D (i) and F E (i) are the flow values ​​at the i-th sampling time, and M is the number of sampling times.

[0087] Step 2: Deep reinforcement learning-model predictive control algorithm application

[0088] Model prediction and optimization control: Based on the model prediction algorithm, the state space model is constructed using the data collected by the sensors:

[0089] x k+1 =Ax k +Bu k +w k

[0090] x k =[T A1 ,T A2 ,T B1 ,T B2 ,T C1 ,T C2 ,F D ,F E ]

[0091] u k =[u flow ,u pump ]

[0092] System Status x k :Including temperature and flow data at various locations in the cooling system, temperature and flow of cooling grooves, cooling channels and water pump impellers.

[0093] Control input u k : Includes cooling water flow and circulation pump speed.

[0094] Process noise w k : represents the random disturbance in the system.

[0095] A: System state matrix: A linear relationship matrix that describes the transfer of system state from time step k to time step k+1, used to predict the change of system state in the absence of control input.

[0096] B: Control matrix: A linear relationship matrix that describes the impact of control input on system state, used to predict the direct impact of control input on system state.

[0097] Prediction model: Model predictive control predicts future temperature changes based on real-time data and historical data and generates control strategies. Predict the state of the next N steps:

[0098]

[0099] in,

[0100] The system state predicted at step k+i

[0101] xk : Current system status

[0102] A: System state matrix

[0103] B: Control input matrix

[0104] u k+i-1-j : Control input

[0105] Optimization goal: Minimize the objective function to calculate the cooling temperature error and control input energy consumption:

[0106]

[0107] J: Optimization objective function: represents the weighted sum of the deviation of the system state from the reference value and the control input energy consumption within the prediction time range (from k to k+N). The goal is to minimize the function value, thereby achieving precise control of the system state and minimization of energy consumption.

[0108] N: Prediction time step: represents the time step predicted by the model predictive control (MPC) algorithm, that is, predicting the next N time steps starting from the current time step k.

[0109] Predicted state: represents the system state predicted at time step k for the next i steps. Predict the temperature and flow data of the cooling tank, cooling channel and water pump impeller at each future time point.

[0110] x ref :Reference temperature: indicates the reference state that the system expects to achieve, that is, the ideal temperature and flow value. The preset ideal temperature and flow of the cooling tank, cooling channel and water pump impeller.

[0111] Q: State weight matrix: a weighted matrix representing the system state error, used to adjust the weights of the impact of different state variables on the optimization target, ensuring that the deviation of key state variables is more sensitive during the optimization process.

[0112] Weighted State Error Sum of Squares: represents the weighted sum of squares of the error between the predicted state and the reference state. The squares of the difference between the predicted temperature and the reference temperature are weighted and summed.

[0113] u k+i :Control input: represents the control input value at time step k+i. Set value of cooling water flow and circulating pump speed.

[0114] R: Control input weight matrix: represents the weighted matrix of control inputs, which is used to adjust the weights of the impact of different control inputs on the optimization target: ensuring that during the optimization process, adjustments to key control inputs (such as cooling water flow) are more sensitive.

[0115] Weighted control input square sum: represents the weighted square of the control input and the square of the cooling water flow and the circulating pump speed, which are weighted and summed.

[0116] Constraints:

[0117]

[0118] u min ≤u k+i ≤u max

[0119] u k+i :Control input: represents the control input value at time step k+i. Set value of cooling water flow and circulating pump speed.

[0120] Predicted state: represents the system state predicted at time step k for the next i steps. Predict the temperature and flow data of the cooling tank, cooling channel and water pump impeller at each future time point.

[0121] x min : Lower limits of temperature and flow rate.

[0122] x max : Upper limit of temperature and flow rate.

[0123] u min : Lower limit of control input.

[0124] u max : Upper limit of control input.

[0125] Deep Reinforcement Learning

[0126] Status definition:

[0127] s t =[T A1 ,T A2 ,T B1 ,T B2 ,T C1 ,T C2 ,F D ,F E ,u flow ,u pump ]

[0128] Status t : Current system temperature, flow and historical control inputs.

[0129] Action definition: The central control module sets the cooling water flow rate and the circulating pump speed, which is achieved by adjusting the cooling water flow control valve and the speed of the circulating pump 410.

[0130] a t=[u` flow ,u` pump ]

[0131] Action a t : Adjust the cooling water flow and circulation pump speed.

[0132] u` flow : Adjustment value of cooling water flow rate.

[0133] u` pump : Adjustment value of circulation pump speed.

[0134] Reward function:

[0135] r t =-||x t -x ref || 2 -λ||u t || 2

[0136] Weight coefficient λ: used to balance temperature error and energy consumption.

[0137] r t : The reward calculated based on the current system state and the control input.

[0138] x t : Current system status.

[0139] x ref : Reference temperature.

[0140] Policy network: Use the deep deterministic policy gradient algorithm to construct the Q value function Q(s t ,a t ;θ):

[0141]

[0142] Q(s t ,a t ; θ): Q value function: represents the state s t Take action a t When , the expected value of future returns estimated by the policy network parameters θ is used to evaluate the value of different actions in the current state, so as to select the optimal action for execution.

[0143] s t :State: represents the state of the system at time step t. The current system temperature, flow rate and historical control input.

[0144] a t :Action represents the action taken by the system at time step t. Adjust the cooling water flow and circulating pump speed.

[0145] r t : Reward means that the system takes action a at time step t t The reward value is calculated based on the difference between the system temperature and the target temperature, and the energy consumption of the control input.

[0146] γ: The discount factor represents the discount factor of future rewards to the current value, which is used to balance immediate rewards and future rewards. It ensures that the strategy is not too far-sighted or short-sighted when considering future rewards.

[0147] The maximum value of future reward: indicates that at the next time step t+1, from state s t+1 At the beginning, the best action a is selected according to the target network parameter θ′ t+1 The maximum Q value that can be obtained. Used to update the current Q value to reflect the impact of future returns.

[0148] θ: Policy network parameters: represent the weight parameters of the policy network. θ: These parameters are continuously adjusted through the training process so that the policy network can accurately estimate the Q value of each state-action pair.

[0149] θ′: Target network parameter: represents the weight parameter of the target Q network, which is used to stabilize the update of the Q value.

[0150] Experience replay: Store and sample samples in the experience replay buffer to update the network.

[0151] D={(s t ,a t ,r t ,s t+1 )}

[0152] D is the experience playback area

[0153] Network update: Use the gradient descent algorithm to update the policy network parameters θ:

[0154]

[0155] θ: Policy network parameters: represent the weight parameters of the policy network. θ: These parameters are continuously adjusted through the training process so that the policy network can accurately estimate the Q value of each state-action pair.

[0156] α: The learning rate represents the step size parameter in the gradient descent algorithm, which is used to control the magnitude of each update. The learning rate determines the speed at which the weight parameters are adjusted. A larger learning rate can speed up convergence but may cause instability; a smaller learning rate leads to stable convergence but slower speed.

[0157] The gradient of the Q-value function represents the Q-value function Q(st ,a t ;θ) with respect to the policy network parameter θ. The gradient represents the direction and rate of change of the Q-value function in the parameter space under the current parameter settings. By calculating the gradient, we can determine how to adjust the parameter θ to minimize the error.

[0158] Deep reinforcement learning-model predictive control combination

[0159] Predictive model correction: Model predictive control uses control strategies optimized by deep reinforcement learning for prediction and correction to form a closed-loop control.

[0160] Real-time control: Deep reinforcement learning learns historical data and real-time data: Trains a deep deterministic policy gradient algorithm to learn historical data and real-time data. Optimizes control strategies to keep temperature and flow within the ideal range.

[0161] Model predictive control for real-time prediction and correction: Model predictive control uses control strategies optimized by deep reinforcement learning for real-time prediction and correction. It realizes intelligent cooling control and ensures that the system operates in the best state.

[0162] According to the prediction and optimization results of deep reinforcement learning and model predictive control, the system parameters are automatically adjusted: 1. Adjust the flow rate of cooling water F D and F E :Accurately control the flow rate by adjusting the control valve. 2. Adjust the running speed of the circulation pump u pump : By adjusting the motor speed of the circulation pump, the pump speed can be controlled. 3. Adjust the refrigeration effect of the refrigeration box: According to the temperature sensor data, adjust the working parameters of the refrigeration box to ensure cooling efficiency.

[0163] Step 3: Intelligent control and automatic adjustment

[0164] Sensor Data Integration:

[0165] As described in step 1, multiple temperature and flow sensors are installed in the cooling system for real-time monitoring. The data of these sensors are transmitted to the central control module in real time through the wireless communication module. The data transmission frequency is set to sample once per second to ensure the real-time and accuracy of the data.

[0166] Analysis results of the deep reinforcement learning-model predictive control algorithm

[0167] Data analysis and prediction:

[0168] Use deep reinforcement learning-model predictive control algorithm to analyze the data collected by sensors to predict future temperature changes and flow requirements. Generate optimized control strategies to ensure efficient and uniform cooling process.

[0169] Deep reinforcement learning optimizes control strategies:

[0170] Through the deep deterministic policy gradient algorithm, historical data and real-time data are learned to optimize the control strategy of cooling water flow and circulating pump speed. Deep reinforcement learning is constantly learning and improving, so that the control strategy of the system is gradually optimized to achieve a more efficient cooling effect.

[0171] Model predictive control real-time prediction and correction:

[0172] Model predictive control is used to make real-time predictions and corrections based on the control strategy optimized by deep reinforcement learning to form a closed-loop control system. Model predictive control uses real-time data to make short-term predictions, correct possible deviations, and ensure the accuracy of the control strategy in actual applications.

[0173] Automatic adjustment mechanism

[0174] Automatic adjustment of cooling water flow:

[0175] According to the prediction and optimization results of the deep reinforcement learning-model predictive control algorithm, the cooling water flow rate is automatically adjusted to ensure efficient and uniform cooling process. According to the current system status and control strategy, the cooling water flow set value is adjusted. The water flow rate is adjusted according to real-time demand to adapt to different cooling needs and improve cooling efficiency.

[0176] Automatic adjustment of circulating pump speed:

[0177] Automatically adjust the operating speed of the circulation pump 410 to ensure that the circulation speed of the cooling water in the cooling tank 403 meets the system requirements. According to the current system status and control strategy, adjust the circulation pump speed setting value. Optimize the fluidity of the cooling water to ensure that the cooling effect reaches the best state.

[0178] Automatic adjustment of refrigeration effect of refrigeration box:

[0179] According to the temperature sensor data, the refrigeration effect of the refrigeration box 413 is automatically adjusted to ensure the maximum cooling efficiency, and the working parameters of the refrigeration box are adjusted according to the current temperature to achieve the best cooling effect. According to the actual temperature requirements, the working mode of the refrigeration box is automatically adjusted to improve the refrigeration efficiency and ensure the stability and uniformity of the cooling process.

[0180] Through intelligent control and automatic regulation, the entire cooling process can be dynamically adjusted according to real-time data and prediction results, improving the system's adaptability and operating efficiency. Using deep reinforcement learning-model predictive control algorithms, the system can continuously optimize and improve control strategies to achieve more intelligent and efficient cooling control to meet the needs of different working conditions.

[0181] The present invention also provides a water pump impeller of a water pump impeller manufacturing device, wherein an impeller mechanism 7 is arranged inside a cooling mechanism 4, the impeller mechanism 7 comprises a rim 71, the rim 71 is arranged in a placement groove of an impeller placement plate 406, a hub 72 is fixedly installed at the middle position of the upper surface of the rim 71, a blade 73 is fixedly installed on the outer side of the hub 72, and an anti-rust blade 74 is fixedly installed on the outer wall of the blade 73.

[0182] Example 4, please refer to the attached Figure 1 -Attached Figure 5 A rust-proof blade 74 is fixedly installed on the blade 73 of the water pump impeller along the direction of the blade 73, which can cut impurities such as fibers in the working environment of the water pump impeller to prevent the fibers from being entangled on the water pump impeller and affecting the operation of the water pump impeller.

[0183] Specifically, the circulation pump 410 draws water into the cooling channel 408, and then filters impurities in the water through the filter element 412 of the filter box 411. The filtered water enters the interior of the refrigeration box 413 for cooling. Finally, the cooled water flows back to the interior of the cooling channel 408 from the water outlet pipe 414, and cools the air in the cooling tank 403, thereby cooling the cast water pump impeller, accelerating the cooling speed, and improving the manufacturing efficiency of the water pump impeller.

[0184] Specifically, the motor 501 rotates to drive the water pump impeller on the turntable 502 to rotate, the air is compressed by the air compressor 507, and then stored in the air tank 508, and then the impurities and moisture in the air are filtered through the filter 509 to ensure the cleanliness of the air for blowing, and then the pressure of the compressed air is adjusted by the pressure regulating valve 510 to meet the blowing demand, and finally, the nozzle 512 at the lower end of the air pipeline 511 sprays gas to blow away the debris on the surface of the water pump impeller, so that the water pump impeller can be blown in all directions without dead ends.

[0185] Specifically, the suction pump 61 sucks the debris in the transparent sealing box 504 from the adapter cover 64, and the debris is adsorbed by the interception filter 65. The gas and fine dust enter the internal air path of the suction pump 61 through the docking tube 63 and are then discharged, which can cooperate with the debris air sweeping mechanism 5 to thoroughly clean the debris on the water pump impeller.

[0186] Specifically, the rust-proof blade 74 on the blade 73 cuts impurities such as fibers in the working environment of the water pump impeller, thereby preventing the fibers from being entangled on the water pump impeller and affecting the operation of the water pump impeller.

[0187] Working principle:

[0188] First, insert the placement rack 404 into the cooling tank 403, place the newly cast water pump impeller in the placement tank of the impeller placement plate 406, then introduce cold water into the cooling channel 408 from the water inlet pipe 402, and the cold water is pumped into the cooling channel 408 from the return pipe 409 by the circulation pump 410, and then enters the filter box 411 from the water outlet end of the circulation pump 410, and the impurities and dirt in the water are filtered by the filter element 412, and the filtered water enters the refrigeration box 413 to be refrigerated, and then flows back from the water outlet pipe 414 to the inside of the cooling channel 408, and the air in the circulating cooling tank 403 of the circulation pump 410 is refrigerated, and the newly cast water pump impeller in the placement rack 404 is cooled, open the cover plate 506, and place the cooled water pump impeller on the turntable 502, and then cover the cover plate 506 on the upper end of the transparent sealing box 504, start the motor 501, so that the water pump impeller moves in a circle along the direction of the turntable 502, The air is compressed by the air compressor 507 and then stored in the air tank 508. Subsequently, the impurities and moisture in the air are filtered through the filter 509 to ensure the cleanliness of the air for purging. The pressure of the compressed air is then adjusted by the pressure regulating valve 510 to meet the purging needs. Finally, the gas is sprayed through the nozzle 512 at the lower end of the air pipeline 511 to purge the debris on the surface of the water pump impeller. When the debris on the surface of the water pump impeller is purged, the suction pump 61 is turned on, the debris is sucked in from the adapter cover 64, and the debris is adsorbed by the interception filter 65. The gas and fine dust enter the internal air path of the suction pump 61 through the docking pipe 63 and are then discharged. When the manufactured water pump impeller is in use, the rust-proof blade 74 on the blade 73 cuts the fibers and other impurities in the working environment of the water pump impeller to prevent the fibers from being entangled in the water pump impeller and affecting the operation of the water pump impeller. At this point, the entire work process ends.

[0189] The above-mentioned front, back, left, right, top and bottom are all based on the figures in the specification. 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.

[0190] 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" and the like indicate orientations or positional relationships based on the orientations or positional relationships 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 should not be understood as limiting the scope of protection of the present invention.

[0191] It should be noted that the device structure and the drawings of the present invention mainly describe the principle of the present invention. In terms of the technology of the design principle, the settings of the power mechanism, power supply system and control system of the device are not fully described. On the premise that the technical personnel in this field understand the principle of the above invention, the details of the power mechanism, power supply system and control system can be clearly known. The control method of the application document is automatic control through a controller, and the control circuit of the controller can be realized by simple programming by the technical personnel in this field.

[0192] The standard parts used therein can all be purchased from the market and can be customized according to the instructions and drawings. The specific connection methods of each part adopt conventional means such as mature bolts, rivets, welding, etc. in the prior art. The machinery, parts and equipment all adopt conventional models in the prior art, and the structures and principles of the components known to technical personnel in this field can be known by these technical personnel through technical manuals or through conventional experimental methods.

[0193] The embodiments of the present invention are described in detail above 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 are made to these embodiments without departing from the principle and spirit of the present invention, and they still fall within the protection scope of the present invention.

Claims

1. A method for manufacturing a water pump impeller of a water pump impeller manufacturing device, the manufacturing device comprising a base (1) and a cooling mechanism (4) arranged on the upper surface of the base near the left side, characterized in that: A debris air sweeping mechanism (5) is provided on the upper surface of the machine base (1) near the upper right position, and a collecting mechanism (6) is provided on the front side of the debris air sweeping mechanism (5); The cooling mechanism (4) comprises a cooling box (401), the cooling box (401) being movably connected to a placement rack (404) via a cooling trough (403), a cooling channel (408) being provided inside the cooling box (401), one end of the cooling channel (408) being fixedly connected to a circulation pump (410) via a return pipe (409), the circulation pump (410) being fixedly connected to a refrigeration box (413) via a filter box (411), and the other end of the cooling channel (408) being fixedly connected to the refrigeration box (413) via a water outlet pipe (414); The debris air sweeping mechanism (5) comprises a motor (501) and a transparent sealed box (504); a rotating disk (502) is fixedly mounted on the output shaft of the motor (501); a slot (505) is provided on the front of the transparent sealed box (504); a cover plate (506) is movably mounted on the upper end of the transparent sealed box (504); an air compressor (507) is fixedly mounted on the upper surface of the cover plate (506) near the left side; the air compressor (507) is fixedly connected to a filter (509) via an air storage tank (508); the filter (509) is fixedly connected to an air pipeline (511) via a pressure regulating valve (510); a nozzle (512) is fixedly mounted on the bottom of the air pipeline (511); The collecting mechanism (6) comprises an air suction pump (61), the air suction pump (61) is fixedly connected to an adaptor cover (64) via a dust suction pipe (62) and a butt joint pipe (63), and an interception filter (65) is inserted into the interior of the adaptor cover (64); An impeller mechanism (7) is arranged inside the cooling mechanism (4), and the impeller mechanism (7) comprises a rim (71), the rim (71) is arranged in a placement groove of an impeller placement plate (406), a hub (72) is fixedly mounted at the middle position of the upper surface of the rim (71), blades (73) are fixedly mounted on the outer side of the hub (72), and rust-proof blades (74) are fixedly mounted on the outer wall of the blades (73); The specific steps of the manufacturing method are as follows: Step 1: First, insert the placement rack (404) into the cooling tank (403), place the newly cast water pump impeller in the placement tank of the impeller placement plate (406), then introduce cold water from the water inlet pipe (402) into the cooling channel (408), the cold water is pumped into the cooling channel (408) from the return pipe (409) by the circulation pump (410), and then enters the filter box (411) from the water outlet of the circulation pump (410), and the impurities and dirt in the water are filtered by the filter element (412), and the filtered water enters the refrigeration box (413) to be refrigerated, and then flows back from the water outlet pipe (414) to the inside of the cooling channel (408), and the air in the circulating cooling tank (403) of the circulation pump (410) is refrigerated, and the newly cast water in the placement rack (404) is cooled. The water pump impeller is cooled; step 2, open the cover plate (506), place the cooled water pump impeller on the turntable (502), then cover the cover plate (506) on the upper end of the transparent sealing box (504), start the motor (501), so that the water pump impeller moves in a circle along the direction of the turntable (502), compress the air through the air compressor (507), and then store it in the air storage tank (508), then filter the impurities and moisture in the air through the filter (509) to ensure the cleanliness of the air for purging, and then adjust the pressure of the compressed air through the pressure regulating valve (510) to meet the purging requirements, and finally spray gas through the nozzle (512) at the lower end of the air pipeline (511) to purge the debris on the surface of the water pump impeller; Step 3: When the debris on the surface of the water pump impeller is purged, the suction pump (61) is turned on, the debris is sucked from the adapter cover (64), the debris is adsorbed by the interception filter (65), and the gas and fine dust enter the internal air path of the suction pump (61) through the docking tube (63) and are then discharged; In step 1, when cooling the water pump impeller, deep reinforcement learning is combined with the intelligent algorithm of the model predictive control algorithm to optimize the cooling process of the water pump impeller and achieve precise cooling control. The specific process is as follows: Step 1: Sensor Layout and Data Collection Temperature sensor: Install temperature sensors at multiple locations of the cooling tank, cooling channel and water pump impeller to monitor the temperature data at each location in real time; Cooling tank: Install a temperature sensor at the top and bottom of the cooling tank, marked as T A1 and T A2 ; Top temperature sensor T A1 : Installed on the top of the cooling tank to monitor the temperature of the air in the cooling tank; Bottom temperature sensor T A2 : Installed at the bottom of the cooling tank to monitor the temperature of the air in the cooling tank; Cooling channel: Install a temperature sensor at the inlet and outlet of the cooling channel, marked as T B1 and T B2 ; Imported temperature sensor T B1 : Installed at the inlet of the cooling channel to monitor the temperature of the cooling water when it enters the cooling channel; Outlet temperature sensor T B2 : Installed at the outlet of the cooling channel to monitor the temperature of the cooling water when it leaves the cooling channel; Impeller: Install a temperature sensor at the center and edge of the pump impeller, marked as T C1 and T C2 ; Core temperature sensor T C1 : Installed in the center of the water pump impeller to monitor the temperature of the impeller center; Edge temperature sensor T C2 : Installed on the edge of the water pump impeller to monitor the temperature of the impeller edge; Flow sensor: Install flow sensors on the water inlet pipe and the water return pipe to monitor the flow of cooling water; Inlet pipe: Install a flow sensor at the inlet of the water inlet pipe, marked as F D ; Used to monitor the cooling water flow entering the cooling channel; Return pipe: Install a flow sensor at the outlet of the return pipe, marked as F E ; Used to monitor the return cooling water flow; Data collection and processing: The data collection module collects data from all sensors to the central control module for real-time data collection and processing; Real-time data transmission The data collected by each temperature sensor and flow sensor is transmitted to the central control module in real time through the wireless communication module; the data transmission frequency is set to sample once per second to ensure real-time monitoring and control; The data acquisition module performs preliminary processing on the collected data, including data verification and format conversion, to ensure the accuracy and consistency of the data; Data processing and analysis The central control module regularly analyzes the collected data and generates time series data of temperature and flow for model prediction and control strategy optimization; Smooth the temperature data to remove noise and outliers; use a moving average filter: in, is the smoothed temperature value, T(ti) is the temperature value at sampling time ti, and N is the window size of the moving average; Perform statistical analysis on the flow data, calculate the average flow and flow change trend; calculate the average flow of the inlet pipe and return pipe: in, and are the average flow rates of the inlet pipe and the return pipe, respectively, and F D (i) and F E (i) are the flow values ​​at the i-th sampling time, and M is the number of sampling times; Step 2: Deep reinforcement learning-model predictive control algorithm application Model prediction and optimization control: Based on the model prediction algorithm, the state space model is constructed using the data collected by the sensors: x k+1 =Ax k +Bu k +w k x k =[T A1 ,T A2 ,T B1 ,T B2 ,T C1 ,T C2 ,F D ,F E ] in k =[in flow ,in pump ] System Status x k :Including temperature and flow data at each location of the cooling system, temperature and flow of cooling tanks, cooling channels and water pump impellers; F D and F E The flow rates of the inlet pipe and the return pipe are Control input u k : Including cooling water flow and circulating pump speed; u flow : Cooling water flow value; u pump : value of circulation pump speed; Process noise w k : represents the random disturbance in the system; A: System state matrix, which describes the linear relationship matrix of the system state from time step k to time step k+1, and is used to predict the change of system state in the absence of control input; B: Control matrix, which describes the linear relationship matrix of the influence of control input on system state, and is used to predict the direct influence of control input on system state; Prediction model: Model predictive control predicts future temperature changes based on real-time data and historical data, generates control strategies, and predicts the state of the next N steps: in, The system state predicted at step k+i x k : Current system status A i : The i-th power of the system state matrix A j : The jth power of the system state matrix u k+i-1-j : Control input i: predicted step index, indicating the number of steps predicted from the current time k; j: cumulative index, used to accumulate the impact of the control input at each step in the prediction process on the future state; j is accumulated from 0 to i-1 to consider the cumulative impact of the control input at each time step on the future state; Optimization goal: Minimize the objective function to calculate the cooling temperature error and control input energy consumption: J: Optimization objective function, which represents the weighted sum of the deviation of the system state from the reference value and the control input energy consumption within the prediction time range; the goal is to minimize the function value, thereby achieving accurate control of the system state and minimization of energy consumption; N: prediction time step, which indicates the time step predicted by the MPC algorithm, i.e., predicting the next N time steps starting from the current time step k; Predicted state, which means the system state predicted at time step k for the next i steps; predicts the temperature and flow data of the cooling tank, cooling channel and water pump impeller at each time point in the future; x ref : Reference temperature, which indicates the reference state that the system expects to achieve, that is, the ideal temperature and flow value; the preset ideal temperature and flow of the cooling tank, cooling channel and water pump impeller; Q: state weight matrix, which represents the weighted matrix of system state error, is used to adjust the weight of the influence of different state variables on the optimization target; ensuring that the deviation of key state variables is more sensitive during the optimization process; The weighted sum of squared state errors represents the weighted sum of squares of the errors between the predicted state and the reference state; the squares of the difference between the predicted temperature and the reference temperature, weighted and summed; u k+i : Control input, indicating the control input value at time step k+i; set value of cooling water flow and circulating pump speed; R: Control input weight matrix, which represents the weighted matrix of control inputs and is used to adjust the weights of the impact of different control inputs on the optimization target, ensuring that the adjustment of key control inputs is more sensitive during the optimization process; The weighted control input square sum represents the weighted square of the control input and the square of the cooling water flow and the circulating pump speed, which are weighted and summed; Constraints: in min in k+i in max u k+i : Control input, indicating the control input value at time step k+i; set value of cooling water flow and circulating pump speed; Predicted state, which means the system state predicted at time step k for the next i steps; predicts the temperature and flow data of the cooling tank, cooling channel and water pump impeller at each time point in the future; x min : Lower limits of temperature and flow rate; x max : Upper limit of temperature and flow rate; u min : The lower limit of the control input; u max : The upper limit of control input; Deep Reinforcement Learning Status definition: s t =[T A1 ,T A2 ,T B1 ,T B2 ,T C1 ,T C2 ,F D ,F E ,u flow ,u pump ] Status t : Current system temperature, flow and historical control input; Action definition: The central control module sets the cooling water flow rate and the circulating pump speed, which is achieved by adjusting the cooling water flow control valve and the circulating pump speed; a t =[u` flow ,u` pump ] Action a t : Adjust cooling water flow and circulation pump speed; u` flow : Adjustment value of cooling water flow; u` pump : Adjustment value of circulation pump speed; Reward function: r t =-||x t -x ref || 2 -λ||u t || 2 Weight coefficient λ: used to balance temperature error and energy consumption; r t : The reward calculated based on the current system state and control input; x t : Current system status; x ref : reference temperature; u t : The control output at time t, the control action generated by the intelligent controller at time t, Policy network: Use the deep deterministic policy gradient algorithm to build the Q value function: Q(s t ,a t ; θ): Q value function, indicating that in state s t Take action a t When , the expected value of future returns estimated by the policy network parameters θ is used to evaluate the value of different actions in the current state, so as to select the optimal action for execution; s t State: represents the state of the system at time step t; the temperature, flow rate and historical control input of the current system; a t : Action represents the action taken by the system at time step t; adjust the cooling water flow and circulating pump speed; r t : Reward means that the system takes action a at time step t t The instant reward obtained after the control is completed; the reward value calculated based on the difference between the system temperature and the target temperature, and the energy consumption of the control input; γ: The discount factor represents the discount coefficient of future rewards to the current value, which is used to balance immediate rewards and future rewards; ensuring that the strategy is not too far-sighted or short-sighted when considering future returns; The maximum value of future reward means that at the next time step t+1, from state s t+1 At the beginning, the best action a is selected according to the target network parameter θ′ t+1 The maximum Q value that can be obtained; used to update the current Q value to reflect the impact of future returns; θ: Policy network parameters, which represent the weight parameters of the policy network. These parameters are continuously adjusted during the training process so that the policy network can accurately estimate the Q value of each state-action pair. θ′: target network parameter, which represents the weight parameter of the target Q network and is used to stabilize the update of the Q value; Experience replay: Store and sample samples in the experience replay buffer to update the network; D=({s t ,a t ,r t ,s t+1 )} D is the experience playback area Network update: Use the gradient descent algorithm to update the policy network parameters θ: θ: Policy network parameters, which represent the weight parameters of the policy network. These parameters are continuously adjusted during the training process so that the policy network can accurately estimate the Q value of each state-action pair. α: The learning rate represents the step size parameter in the gradient descent algorithm, which is used to control the magnitude of each update; the learning rate determines the speed of weight parameter adjustment; a larger learning rate can speed up convergence but may cause instability; a smaller learning rate leads to stable convergence but slower speed; ▽ θ Q(s t ,a t ; θ): The gradient of the Q-value function represents the Q-value function Q(s t ,a t ; θ) is the gradient of the policy network parameter θ; the gradient represents the direction and rate of change of the Q-value function in the parameter space under the current parameter setting; by calculating the gradient, we can determine how to adjust the parameter θ to minimize the error; Deep reinforcement learning-model predictive control combination Predictive model correction: Model predictive control uses control strategies optimized by deep reinforcement learning to make predictions and corrections, forming a closed-loop control; Real-time control: Deep reinforcement learning learns historical data and real-time data, trains deep deterministic policy gradient algorithms, learns historical data and real-time data; optimizes control strategies to keep temperature and flow within ideal ranges; Model predictive control for real-time prediction and correction: Model predictive control uses control strategies optimized by deep reinforcement learning for real-time prediction and correction; it implements intelligent cooling control to ensure that the system operates in the best condition; According to the prediction and optimization results of deep reinforcement learning and model predictive control, the system parameters are automatically adjusted:

1. Adjust the flow rate of cooling water F D and F E : Accurately control the flow rate by adjusting the control valve; 2. Adjust the running speed of the circulation pump u pump : By adjusting the motor speed of the circulation pump, the pump speed can be controlled; 3. Adjust the refrigeration effect of the refrigeration box: According to the temperature sensor data, adjust the working parameters of the refrigeration box to ensure cooling efficiency; Step 3: Smart Control and Automatic Setup Sensor data integration: Multiple temperature and flow sensors are installed in the refrigeration system to monitor the operating status of the system in real time. The data from all sensors are transmitted to the central controller through the wireless communication module. The data transmission rate is set to once per second to ensure the accuracy and veracity of real-time data; According to the prediction results and optimization of the deep gain model predictive control algorithm, the cooling water flow rate is automatically adjusted to ensure the efficiency and uniformity of the cooling process. According to the current state of the system and the control strategy, the set value of the cooling water flow rate is adjusted to adapt to different cooling needs and improve cooling efficiency; Automatically adjust the running speed of the circulation pump to ensure that the circulation speed of the cooling water in the cooling tank meets the system requirements. According to the current system status and control strategy, adjust the circulation pump speed setting value, optimize the cooling water flow, and ensure that the cooling power reaches the optimal level; According to the temperature sensor data, the cooling power of the refrigeration box can be automatically adjusted to ensure maximum cooling efficiency; the working parameters of the refrigeration box are adjusted according to the current temperature, and the operation mode of the refrigeration box is automatically adjusted according to the actual temperature requirements to ensure the stability and consistency of the cooling process.

2. The method according to claim 1, characterized in that: A mounting groove (2) is provided inside the machine base (1) near the upper right position, and a plug-in groove (3) is provided outside the mounting groove (2).

3. The method according to claim 2, characterized in that: The bottom of the cooling box (401) is fixedly mounted on the upper surface of the machine base (1) near the upper left position, a sealing cover (407) is movably mounted on the upper end of the cooling box (401), the cooling groove (403) is opened inside the cooling box (401), the placement rack (404) is plugged into the inside of the cooling groove (403), an air vent (405) is opened on the outer wall of the placement rack (404), an impeller placement plate (406) is arranged inside the placement rack (404), and a water inlet pipe (402) is fixedly mounted on the left side of the cooling box (401), and the water inlet pipe (402) is communicated with a cooling channel (408).

4. The method according to claim 3, characterized in that: One end of the return water pipe (409) is fixedly mounted on one end of the cooling channel (408), the other end of the return water pipe (409) is fixedly connected to the return water end of the circulation pump (410), one end of the filter box (411) is fixedly connected to the water outlet end of the circulation pump (410), a filter element (412) is inserted into the filter box (411), one end of the refrigeration box (413) is fixedly connected to the other end of the filter box (411), the other end of the refrigeration box (413) is fixedly connected to one end of the water outlet pipe (414), and the other end of the water outlet pipe (414) is fixedly mounted on the other end of the cooling channel (408).

5. The method according to claim 2, characterized in that: The bottom of the motor (501) is fixedly mounted inside the mounting groove (2), and a plug-in component (503) is fixedly mounted on the bottom of the transparent sealing box (504), and the plug-in component (503) is plugged into the inside of the plug-in groove (3).

6. The method according to claim 5, characterized in that: The air storage tank (508) is fixedly connected to the air outlet pipe of the air compressor (507), the filter (509) is fixedly connected to the air outlet pipe of the air storage tank (508), the pressure regulating valve (510) is fixedly installed on the air outlet pipe of the filter (509), the upper end of the air delivery pipeline (511) is fixedly connected to the pressure regulating valve (510), and the nozzle (512) is located inside the transparent sealing box (504) near the upper end.

7. The method according to claim 2, characterized in that: The collecting mechanism (6) comprises an air suction pump (61), the air suction pump (61) is fixedly mounted on the upper surface of the machine base (1) near the lower right position, one end of the dust suction pipe (62) is fixedly mounted on the suction end of the air suction pump (61), one end of the butt joint pipe (63) is fixedly connected to the other end of the dust suction pipe (62), the other end of the butt joint pipe (63) is fixedly connected to one end of the adapter cover (64), and the other end of the adapter cover (64) is fixedly connected to the slot (505).

Images