Temperature prediction and visualization method for intelligent wireless vacuum furnace

Abstract

The invention provides a temperature prediction and visualization method for an intelligent wireless vacuum furnace, and relates to the technical field of vacuum furnaces. A vacuum gauge mounting hole is formed outside a furnace body of the intelligent wireless vacuum furnace and used for mounting a vacuum degree measuring device and a split type wireless passive temperature sensor; through wireless passive temperature sensors and current and voltage sensors, voltage, current and temperature data and out-of-furnace vacuum degree of eight independent partitions are acquired in real time, space temperature and power data are monitored at different working moments, and compared with single-point temperature measurement, the temperature characterization precision of the partitions can be improved; and real-time data support is provided for prediction of a subsequent network model. According to the method, a prediction method combining ANSYS simulation and a CNN-LSTM deep learning model is provided, furnace temperature distribution results under different furnace charging modes are predicted in real time, real-time data results are displayed on a screen in a real-time cloud picture conversion mode through a data conversion image processing technology, and the furnace temperature is monitored in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of vacuum furnaces, and more specifically, to a temperature prediction and visualization method for an intelligent wireless vacuum furnace. Background Art

[0002] Vacuum furnaces are widely used in the fields of metal heat treatment, ceramic sintering, brazing, etc. due to their characteristics of non-oxidizing working environment and high controllability of uniform temperature.

[0003] Currently, when heat-treating silicon carbide workpieces in an extra-large vacuum sintering furnace, due to the large size of the furnace body, eight independent heaters are required to perform zoning heating inside the furnace, and there are significant differences in temperature among different zones. At present, the temperature inside the furnace is mainly monitored by installing temperature sensors at various parts of the vacuum furnace for temperature measurement. However, due to the high final sintering temperature inside the furnace, real-time temperature measurement cannot be achieved by arranging sensors inside the furnace, and the temperature measured by the sensors only represents the local temperature, and the real-time change of the entire temperature field inside the furnace cannot be intuitively observed. Moreover, it is also easily restricted by problems such as the number and position of sensors and the complexity of wiring. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the purpose of the present invention is to propose a temperature prediction and visualization method for an intelligent wireless vacuum furnace, including:

[0005] Step 1: Construct an intelligent wireless vacuum furnace, which includes 8 single-group furnace cavities. Each single-group furnace cavity includes a furnace shell, carbon felt, graphite plate, heating element group, power connection device, vacuum measurement device, and multiple split-type wireless passive temperature sensors. Among them, the power connection device includes wireless voltage and current measurement sensors, and the vacuum measurement device includes a three-way valve port high-precision wireless temperature sensor and a vacuum gauge measurement sensor;

[0006] Step 2: Place a temperature measurement ring at a preset cross-section in each single-group furnace cavity. Run the intelligent wireless vacuum furnace. During the operation process, for each single-group furnace cavity, continuously collect the voltage measurement value of the intelligent wireless vacuum furnace and the current measurement value of the intelligent wireless vacuum furnace through the wireless voltage and current measurement sensors; continuously measure the first temperature of the vacuum measurement device through the three-way valve port high-precision wireless temperature sensor, and continuously measure the external vacuum measurement value of the intelligent wireless vacuum furnace through the vacuum measurement device; continuously collect multiple second temperatures through the split-type wireless passive temperature sensors; collect the deformation amount of the temperature measurement ring after the operation of the intelligent wireless vacuum furnace ends;

[0007] Step 3: At each running time during the operation, for each single-group furnace chamber, calculate the power based on the voltage measurement value and the current measurement value, calculate the in-furnace vacuum degree based on the first temperature, multiple second temperatures, and the out-of-furnace vacuum degree measurement value. The power, in-furnace vacuum degree, and multiple second temperatures form target data. Each single-group furnace chamber corresponds to one target data. Normalize each target data to obtain the normalized target data for each single-group furnace chamber, and further obtain the normalized target data for each single-group furnace chamber corresponding to each running time. Determine the standard temperature through the deformation amount of the temperature measurement ring;

[0008] Step 4: Construct a three-dimensional model of the intelligent wireless vacuum furnace. Based on the normalized target data for each single-group furnace chamber corresponding to each running time and the three-dimensional model of the intelligent wireless vacuum furnace, perform simulation through simulation software to obtain the temperature distribution data for each preset cross-section corresponding to each running time. For each preset cross-section, correct the temperature distribution data based on the standard temperature and multiple second temperatures to obtain the corrected temperature distribution data, and further obtain the corrected temperature distribution data for each preset cross-section. Normalize the corrected temperature distribution data for each preset cross-section to obtain the normalized corrected temperature distribution data for each preset cross-section;

[0009] Step 5: For each running time, use the running time and the normalized target data for each single-group furnace chamber corresponding to this running time as input samples, and use the normalized corrected temperature distribution data for each preset cross-section corresponding to this running time as output samples. The input samples and the output samples form training samples. Each running time corresponds to one training sample, and further obtain multiple training samples. The multiple training samples form a data set. Divide the data set into a training set and a test set according to a preset ratio;

[0010] Step 6: Train the CNN-LSTM prediction model based on the multiple training samples in the training set to obtain the trained CNN-LSTM prediction model;

[0011] Step 7: For different workpiece sizes and materials of the intelligent wireless vacuum furnace, execute Steps 2-6 to obtain multiple trained CNN-LSTM prediction models under different workpiece sizes and materials;

[0012] Step 8: Operate the intelligent wireless vacuum furnace to be predicted, and obtain the normalized data to be predicted for each single-group furnace chamber corresponding to each running time through the wireless voltage and current measurement sensors, three-way valve port high-precision wireless temperature sensors, vacuum degree measurement device, and split-type wireless passive temperature sensors of the intelligent wireless vacuum furnace;

[0013] Step 9: According to the workpiece size and material of the intelligent wireless vacuum furnace to be predicted, obtain the trained CNN-LSTM prediction model corresponding to the workpiece size and material. Input the running time and the normalized data to be predicted for each single furnace cavity at this running time into the obtained trained CNN-LSTM prediction model, and output the temperature distribution data of the preset cross-section in each single furnace cavity of the intelligent wireless vacuum furnace to be predicted. Denormalize the temperature distribution data of each preset cross-section to obtain the denormalized temperature distribution data of each preset cross-section;

[0014] Step 10: Visualize and output the denormalized temperature distribution data of each preset cross-section.

[0015] Optionally, the furnace shell of the single furnace cavity is connected to the power connection device. In the order from outside to inside, the single furnace cavity includes carbon felt, graphite plate, and heating element group from the furnace shell inward; the furnace shell of the single furnace cavity is provided with wireless passive temperature sensor mounting holes, and the wireless passive temperature sensor mounting holes include left mounting holes, top mounting holes, and right mounting holes. The furnace shell of the single furnace cavity includes four planes, namely the first plane, the second plane, the third plane, and the fourth plane. The first plane and the third plane are parallel, the second plane and the fourth plane are parallel, the second plane is perpendicular to both the first plane and the third plane. The first plane is provided with two left mounting holes, the second plane is provided with two top mounting holes, the third plane is provided with two right mounting holes, ceramic fixing nuts are fixed above the left mounting holes, top mounting holes, and right mounting holes, and the third plane is provided with a vacuum gauge mounting hole.

[0016] Optionally, the vacuum measurement device includes a detachable vacuum gauge port pipe, a three-layer honeycomb filter molybdenum plate, an in-furnace connecting molybdenum pipe, an out-of-furnace water-cooled pipe, a three-way valve, a high-precision wireless temperature sensing and vacuum gauge measurement sensor at the three-way valve port; the three-layer honeycomb filter molybdenum plate includes a first-layer honeycomb filter molybdenum plate, a second-layer honeycomb filter molybdenum plate, and a third-layer honeycomb filter molybdenum plate. Among them, the pore size of the internal holes in each layer of the honeycomb filter molybdenum plate is the same, and the pore size of the internal holes in the first-layer honeycomb filter molybdenum plate, the pore size of the internal holes in the second-layer honeycomb filter molybdenum plate, and the pore size of the internal holes in the third-layer honeycomb filter molybdenum plate are all different;

[0017] The out-of-furnace water-cooled pipe is provided with a water-cooled pipe water inlet and a water-cooled pipe water outlet. Flange plates are provided at both ends of the out-of-furnace water-cooled pipe. The three-way valve is provided with three ports, namely the first port, the second port, and the third port. Flange plates are provided at the first port, the second port, and the third port;

[0018] The detachable vacuum gauge port tube is fixed at the vacuum gauge mounting hole through a flange; the three-layer honeycomb-shaped filtering molybdenum plate is embedded in the detachable vacuum gauge port tube, the detachable vacuum gauge port tube is embedded in the first end of the in-furnace connecting molybdenum tube, a flange is provided at the second end of the in-furnace connecting molybdenum tube, the flange at the second end of the in-furnace connecting molybdenum tube is connected to the flange at one end of the out-of-furnace water-cooled pipe, the flange at the other end of the out-of-furnace water-cooled pipe is connected to the flange at the first port of the three-way valve, a flange is provided at the port of the high-precision wireless temperature sensor at the three-way valve port, the flange at the second port of the three-way valve is connected to the flange at the port of the high-precision wireless temperature sensor at the three-way valve port, a flange is provided at the port of the vacuum gauge measurement sensor, and the flange at the third port of the three-way valve is connected to the flange at the port of the vacuum gauge measurement sensor.

[0019] Optionally, the split-type wireless passive temperature sensor includes an external energy supply and signal processing module, a waveguide channel, a front-end probe sensor module, and a silicon carbide protective housing;

[0020] The front-end probe sensor module includes an alumina protective film, a diamond film, interdigital electrodes, and a sapphire substrate. The interdigital electrodes are placed on the sapphire substrate, the diamond film is deposited on the interdigital electrodes and the sapphire substrate, and the alumina protective film is deposited on the diamond film;

[0021] The external energy supply and signal processing module includes an antenna system, a signal processing module, a signal transmission module, and a thermoelectric energy supply module. The thermoelectric energy supply module is a high-performance thermoelectric material;

[0022] The silicon carbide protective housing is cylindrical and includes a first end and a second end. The antenna system is connected to the first end of the silicon carbide protective housing. A signal processing module is provided inside the first end of the silicon carbide protective housing. The front-end probe sensor module is embedded inside the second end of the silicon carbide protective housing. In the direction from outside to inside, a strip-shaped thermoelectric energy supply module, a cylindrical waveguide channel, and a signal transmission module are arranged from the silicon carbide protective housing inward. The thermoelectric energy supply module is connected to the signal processing module. One end of the waveguide channel is connected to the signal processing module, the other end of the waveguide channel is connected to the front-end probe sensor module, and the signal transmission module is connected to the signal processing module;

[0023] External threads are provided on the outside of the second end of the silicon carbide protective housing, and the external threads at the second end of the silicon carbide protective housing are engaged with the internal threads of the ceramic fixing nut.

[0024] Optionally, in step 3, based on the first temperature, multiple second temperatures, and the measured value of the out-of-furnace vacuum degree, calculating the in-furnace vacuum degree includes:

[0025] Step A1: Among multiple second temperatures, determine the second temperatures collected by two split-type wireless passive temperature sensors fixed on the third plane of the furnace shell, and calculate the average temperature of the second temperatures collected by the two split-type wireless passive temperature sensors on the third plane of the furnace shell;

[0026] Step A2: Based on the average temperature, the first temperature, and the measured value of the external furnace vacuum degree, calculate the internal furnace vacuum degree, which is specifically achieved through the following formula:

[0027]

[0028] where, T1` is the first temperature, T2` is the average temperature, T1 and T2 are preset calibration temperatures, P2 is the measured external furnace vacuum degree, and P2` is the internal furnace vacuum degree.

[0029] Optionally, step 6 specifically includes:

[0030] Input the input sample in the training sample into the CNN-LSTM prediction model to obtain predicted temperature distribution data. According to the predicted temperature distribution data and the corrected temperature distribution data, calculate the loss function, and determine whether the loss function is greater than the preset threshold. When the loss function is not greater than the preset threshold, obtain the trained CNN-LSTM prediction model. When the loss function is greater than the preset threshold, update the parameters of the CNN-LSTM prediction model based on the loss function, and obtain the next training sample, then return and execute: input the input sample in the training sample into the CNN-LSTM prediction model.

[0031] Optionally, step 8 specifically includes:

[0032] Continuously collect the target voltage value of the intelligent wireless vacuum furnace to be predicted and the target current value of the intelligent wireless vacuum furnace to be predicted through a wireless voltage and current measurement sensor; continuously measure the first target temperature of the vacuum degree measurement device through a three-way valve port high-precision wireless temperature sensor, and continuously measure the target external furnace vacuum degree of the intelligent wireless vacuum furnace through the vacuum degree measurement device; continuously collect multiple second target temperatures through split-type wireless passive temperature sensors;

[0033] At each running time during the operation, for each single group of furnace cavities, calculate the target power based on the target voltage value and the target current value, calculate the target internal furnace vacuum degree based on the first target temperature, multiple second target temperatures, and the target external furnace vacuum degree. The target power, the target external furnace vacuum degree, and the multiple second target temperatures form the data to be predicted. Normalize each data to be predicted. Each single group of furnace cavities corresponds to one data to be predicted, and obtain the normalized data to be predicted, and further obtain the normalized data to be predicted for each single group of furnace cavities corresponding to each running time.

[0034] The beneficial effects of adopting the above technical solutions are as follows:

[0035] In the present invention, a vacuum gauge mounting hole is provided at the middle position of the side of eight independent partitions, and an ionization gauge vacuum gauge is installed to measure the vacuum degree of each independent partition of the vacuum sintering furnace respectively. The connection between the vacuum gauge tube and the sintering furnace chamber body is cooled by a water-cooled pipe, and its length is appropriately increased to solve the problems of large fluctuations in the indication of the vacuum gauge and relatively serious errors. The installation of the external ionization gauge vacuum gauge adopts a three-way structure design. While an ionization gauge vacuum gauge is installed at the three-way valve, a temperature sensor is also installed to monitor the temperature T1ˋ of the control gauge tube. By monitoring and controlling the cooling rate of the water-cooling device, the working environment is kept stable near the calibration temperature T1, thereby reducing the measurement error and possible false responses and improving the accuracy of vacuum measurement of the vacuum sintering furnace. The measurement port of the ionization gauge vacuum gauge in the furnace chamber adopts a design of multiple-layer honeycomb-shaped molybdenum plates embedded in the port tube, which reduces heat radiation loss, filters the coke gas in the furnace, reduces the interference to the uniformity of the temperature field in the furnace and is convenient for replacement; the wireless passive temperature sensor is connected through a ceramic nut fixed on the furnace shell, which is convenient for disassembly and replacement while being fixed; the wireless passive temperature sensor adopts a split layout, which enhances the signal penetration ability and reduces the working temperature of the signal acquisition module; the hot spot power supply module of the wireless passive temperature sensor adopts a longitudinal structure layout, and generates electric energy by means of a significant longitudinal temperature difference. Through the wireless passive temperature sensor and the current-voltage sensor, the real-time input voltage, current and temperature data of the top, left and right sides of the furnace chamber of eight independent partitions are extracted, which solves the monitoring of space temperature and power data at different working times. Compared with single-point temperature measurement, the characterization accuracy of the partition temperature can be improved, providing real-time data support for the prediction of the subsequent network model. A prediction method combining ANSYS simulation and CNN-LSTM deep learning model is proposed to predict the temperature distribution results in the furnace under different furnace loading methods in real time, and the real-time data results are displayed on the screen in the form of a real-time transformed cloud map through data conversion image processing technology to monitor the temperature in the furnace in real time; the Kriging spatial interpolation method is selected to dynamically adjust the resolution of the temperature cloud map, which can reduce the amount of data required for prediction and the prediction time. Description of the Drawings

[0036] Figure 1 It is a schematic flow chart of a temperature prediction and visualization method for an intelligent wireless vacuum furnace in an embodiment of the present invention;

[0037] Figure 2 It is the front view of a single group of furnace chambers in an embodiment of the present invention;

[0038] Figure 3 It is the perspective view of the structure of the heating element group in an embodiment of the present invention;

[0039] Figure 4It is the perspective view of the single-group furnace chamber graphite heating element in the embodiment of the present invention;

[0040] Figure 5 It is the top view of the single-group furnace chamber in the embodiment of the present invention;

[0041] Figure 6 It is the overall side view of the intelligent wireless vacuum furnace in the embodiment of the present invention;

[0042] Figure 7 It is the schematic diagram of the water-cooled electrode structure in the embodiment of the present invention;

[0043] Figure 8 It is the side view of the single-group furnace chamber mounting hole structure in the embodiment of the present invention;

[0044] Figure 9 It is the overall structure schematic diagram of the vacuum degree measuring device in the embodiment of the present invention;

[0045] Figure 10 It is the disassembled structure schematic diagram of the vacuum degree measuring device in the embodiment of the present invention;

[0046] Figure 11 It is the disassembled structure schematic diagram of the coke gas filtering device of the vacuum degree measuring device in the embodiment of the present invention;

[0047] Figure 12 It is the installation structure schematic diagram of the split-type wireless passive temperature sensor and the vacuum degree measuring device in the embodiment of the present invention;

[0048] Figure 13 It is the cross-sectional view of the split-type wireless passive temperature sensor in the embodiment of the present invention;

[0049] Figure 14 It is the exploded view of the front-end probe sensor module of the split-type wireless passive temperature sensor in the embodiment of the present invention;

[0050] Figure 15 It is the detailed cross-sectional view of the structure of the split-type wireless passive temperature sensor in the embodiment of the present invention;

[0051] Figure 16 It is the vacuum furnace temperature prediction and visualization flow chart of the CNN-LSTM neural network model in the embodiment of the present invention;

[0052] Figure 17 It is the distribution map of the temperature prediction points of the cross-section of the vacuum sintering furnace in the embodiment of the present invention;

[0053] Figure 18 It is the display diagram of the furnace temperature prediction and visualization platform for each zone of the vacuum sintering furnace in the embodiment of the present invention;

[0054] In the figure: 1 - furnace shell, 101 - installation hole for wireless passive temperature sensor, 1001 - left installation hole, 1002 - top installation hole, 1003 - right installation hole, 102 - ceramic fixing nut, 103 - installation hole for vacuum gauge; 2 - carbon felt; 3 - graphite plate; 4 - heating element group; 401 - graphite heating tube, 402 - power connection plate; 4021 - power connection plate A, 4022 - power connection plate B, 4023 - power connection plate C; 403 - port fixing nut, 404 - power-on straight plate; 4041 - bottom power-on straight plate, 4042 - side power-on straight plate, 4043 - corner power-on plate; 5 - power connection device; 501 - water-cooled electrode; 5011 - A pole of water-cooled electrode, 5012 - B pole of water-cooled electrode, 5013 - C pole of water-cooled electrode, 5014 - water inlet of water-cooled electrode, 5015 - water outlet of water-cooled electrode; 502 - three-phase electrode power transmission plate; 5021 - three-phase electrode power transmission plate A, 5022 - three-phase electrode power transmission plate B, 5023 - three-phase electrode power transmission plate C; 503 - electrode fixing flange, 504 - wireless voltage and current measurement sensor; 6 - vacuum degree measurement device; 601 - detachable vacuum gauge port pipe, 602 - three-layer honeycomb filter molybdenum plate, 603 - in-furnace connecting molybdenum pipe, 604 - out-of-furnace water-cooled pipe; 6041 - water inlet of water-cooled pipe, 6042 - water outlet of water-cooled pipe, 605 - three-way valve, 606 - high-precision wireless temperature sensor at three-way valve port, 607 - vacuum gauge measurement sensor; 7 - split-type wireless passive temperature sensor; 701 - external energy supply and signal processing module; 7011 - antenna system, 7012 - signal processing module, 7013 - signal transmitting module, 7014 - thermoelectric energy supply module; 702 - waveguide channel, 703 - front-end probe sensor module; 7031 - alumina protective film, 7032 - diamond thin film, 7033 - interdigital electrode, 7034 - sapphire substrate; 704 - silicon carbide protective shell. Specific embodiments

[0055] The following combines the accompanying drawings and embodiments to further describe in detail the specific embodiments of the present invention. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0056] Aiming at the problems existing in the prior art, the present invention designs a wireless data acquisition module to reduce the limitations of problems such as complex wiring, obtains the operating condition parameters during operation to predict the in-furnace temperature in real time, and displays the real-time cloud map through a computer visualization interface. At the same time, according to the prediction results, a smart and visual temperature prediction system for negative feedback adjustment of the power input in the non-uniform area is provided, so as to improve the overall uniformity of the furnace temperature in eight zones and the sintering quality of workpieces, and provide the operator with real-time and visual in-furnace conditions.

[0057] Specifically, the present invention provides a temperature prediction and visualization method for an intelligent wireless vacuum furnace, combined withFigure 1 , may include the following steps: including:

[0058] Step 1: Construct an intelligent wireless vacuum furnace, the intelligent wireless vacuum furnace includes 8 single-group furnace cavities, the single-group furnace cavity includes a furnace shell 1, carbon felt 2, graphite plate 3, heating element group 4, power connection device 5, vacuum degree measurement device 6 and multiple split-type wireless passive temperature sensors 7. Among them, the power connection device 5 includes a wireless voltage and current measurement sensor 504, and the vacuum degree measurement device 6 includes a three-way valve port high-precision wireless temperature sensor 606 and a vacuum gauge measurement sensor 607;

[0059] Combined with Figure 2 , the furnace shell 1 of the single-group furnace cavity is connected to the power connection device 5. In the order from outside to inside, the single-group furnace cavity from the furnace shell inwards is carbon felt 2, graphite plate 3, and heating element group 4 respectively; Combined with Figure 3 and Figure 4 , the heating element group 4 includes graphite heating tubes 401, power connection plates 402, port fixing nuts 403 and power-on straight plates 404. Among them, the power connection plate 402 includes a power connection plate A 4021, a power connection plate B 4022 and a power connection plate C 4023, and the power-on straight plate 404 includes a bottom power-on straight plate 4041, a side power-on straight plate 4042 and a corner power-on plate 4043; The graphite heating tubes 401 are fixed on the power connection plate 402 and the power-on straight plate 404 by the port fixing nuts 403;

[0060] Among them, the graphite heating tube 401 is the core heating element in the vacuum furnace. The working temperature of the graphite heating tube 401 is between 80 and 2000 degrees Celsius. It converts electrical energy into heat energy through the principle of resistance heating. It is made of high-purity graphite material, has excellent high-temperature resistance performance and good electrical conductivity, and can directly heat the materials in the furnace to provide the required temperature environment.

[0061] Combined with Figure 4 , Figure 5 , Figure 6 and Figure 7 , the power connection device 5 includes; water-cooled electrodes 501, three-phase electrode power transmission plates 502, electrode fixing flanges 503 and wireless voltage and current measurement sensors 504. Among them, the water-cooled electrodes 501 include a water-cooled electrode A pole 5011, a water-cooled electrode B pole 5012, a water-cooled electrode C pole 5013, a water-cooled electrode water inlet 5014 and a water-cooled electrode water outlet 5015; The three-phase electrode power transmission plates 502 include a three-phase electrode power transmission plate A 5021, a three-phase electrode power transmission plate B 5022 and a three-phase electrode power transmission plate C5023; Among them, the wireless voltage and current measurement sensor 504 is located on the three-phase electrode power transmission plate 502.

[0062] Among them, the heating element group 4 adopts a triangular power connection method and is connected to an external power supply through a power connection device 5. The power connection device 5 is used for the energized heating of the heating element group. The power connection plate 402 is connected to the water-cooled electrode 501 for supplying power to the heating element group. The power connection steps are as follows: the three-phase electrode power transmission plate 502 is fixed by connecting with the electrode fixing flange 503 to energize the water-cooled electrode 501, and the water-cooled electrode 501 is connected to the power connection plate 402 in the heating element group 4 to be energized. Among them, the three-phase electrode power transmission plate A 5021 energizes the water-cooled electrode A 5011, and the water-cooled electrode A 5011 is connected to the power connection plate A 4021 in the heating element group 4 to be energized; the three-phase electrode power transmission plate B 5022 energizes the water-cooled electrode B 5012, and the water-cooled electrode B 5012 is connected to the power connection plate B 4022 in the heating element group 4 to be energized; the three-phase electrode power transmission plate C 5023 energizes the water-cooled electrode C 5013, and the water-cooled electrode C 5013 is connected to the power connection plate C 4023 in the heating element group 4 to be energized; the water-cooled electrode 501 inputs cooling water through the water-cooled electrode water inlet 5014 and outputs cooling water at the water-cooled electrode water outlet 5015.

[0063] Combined with Figure 8 , the furnace shell 1 of the single-group furnace chamber is provided with wireless passive temperature sensor mounting holes 101. The wireless passive temperature sensor mounting holes include a left mounting hole 1001, a top mounting hole 1002, and a right mounting hole 1003. The furnace shell 1 of the single-group furnace chamber includes four planes, namely the first plane, the second plane, the third plane, and the fourth plane. The first plane and the third plane are parallel, the second plane and the fourth plane are parallel, the second plane is perpendicular to both the first plane and the third plane. The first plane is provided with two left mounting holes 1001, the second plane is provided with two top mounting holes 1002, the third plane is provided with two right mounting holes 1003. Ceramic fixing nuts 102 are fixed above the left mounting hole 1001, the top mounting hole 1002, and the right mounting hole 1003. The third plane is provided with a vacuum gauge mounting hole 103.

[0064] Combined with Figure 9 , 10 , 11, the vacuum degree measuring device includes a detachable vacuum gauge port pipe 601, a three-layer honeycomb-shaped filter molybdenum plate 602, a furnace internal connecting molybdenum pipe 603, a furnace external water-cooled pipe 604, a three-way valve 605, a three-way valve port high-precision wireless temperature sensor 606, and a vacuum gauge measuring sensor 607;

[0065] The three-layer honeycomb filter molybdenum plate 602 includes a first-layer honeycomb filter molybdenum plate, a second-layer honeycomb filter molybdenum plate, and a third-layer honeycomb filter molybdenum plate. Among them, the pore size of the internal pores in each layer of the honeycomb filter molybdenum plate is the same, and the pore size of the internal pores in the first-layer honeycomb filter molybdenum plate, the pore size of the internal pores in the second-layer honeycomb filter molybdenum plate, and the pore size of the internal pores in the third-layer honeycomb filter molybdenum plate are all different. This design is to isolate the colloidal coke gas in the furnace body from entering the vacuum measurement system and to be disassembled and cleaned after the sintering process is completed.

[0066] The external furnace water-cooled pipe 604 is provided with a water-cooled pipe water inlet 6041 and a water-cooled pipe water outlet 6042. Cooling water is input through the water-cooled pipe water inlet 6041, and cooling water is output through the water-cooled pipe water outlet 6042 for cooling the water-cooled pipe to keep the gas in the pipe within the working range of the vacuum gauge. Flange plates are provided at both ends of the external furnace water-cooled pipe 604. The three-way valve 605 has three ports, namely the first port, the second port, and the third port, and flange plates are provided at the first port, the second port, and the third port.

[0067] Combined with Figure 12 , the detachable vacuum gauge port 601 pipe is fixed to the vacuum gauge mounting hole 103 through a flange plate. The three-layer honeycomb filter molybdenum plate 602 is embedded in the detachable vacuum gauge port pipe 601. The pore sizes of the three-layer honeycomb molybdenum plate filter holes of the three-layer honeycomb filter molybdenum plate 602 are different, which is used to reduce heat radiation loss and filter the coke gas in the furnace, reduce the interference with the uniformity of the temperature field in the furnace and facilitate replacement. The detachable vacuum gauge port pipe 601 is embedded in the first end of the internal furnace connecting molybdenum pipe 603. The molybdenum material design can withstand high temperatures. The pipe orifice of the detachable vacuum gauge port pipe 601 is in contact with the graphite plate 3 in the furnace cavity, and it can be regularly replaced and disassembled and cleaned from the furnace cavity.

[0068] A flange plate is provided at the second end of the internal furnace connecting molybdenum pipe 603. The flange plate at the second end of the internal furnace connecting molybdenum pipe 603 is connected to the flange plate at one end of the external furnace water-cooled pipe 604. The flange plate at the other end of the external furnace water-cooled pipe 604 is connected to the flange plate at the first port of the three-way valve 605. A flange plate is provided at the port of the three-way valve port high-precision wireless temperature sensor 606. The flange plate at the second port of the three-way valve 605 is connected to the flange plate at the port of the three-way valve port high-precision wireless temperature sensor 606. A flange plate is provided at the port of the vacuum gauge measurement sensor 607. The flange plate at the third port of the three-way valve 605 is connected to the flange plate at the port of the vacuum gauge measurement sensor 607.

[0069] Combined with Figure 13 , Figure 14 and Figure 15, the split-type wireless passive temperature sensor 7 includes an external energy supply and signal processing module 701, a waveguide channel 702, a front-end probe sensor module 703, and a silicon carbide protective housing 704. Its structure places the external energy supply and signal processing module that is not resistant to high temperatures outside the furnace.

[0070] The front-end probe sensor module 703 includes an alumina protective film 7031, a diamond thin film 7032, interdigital electrodes 7033, and a sapphire substrate 7034. The interdigital electrodes 7033 are placed on the sapphire substrate 7034. The diamond thin film 7032 is deposited on the interdigital electrodes 7033 and the sapphire substrate 7034. The alumina protective film 7031 is deposited on the diamond thin film 7032; it is used to improve the wear resistance and chemical stability of the front-end probe sensor module 703; the outside of the sensor is a silicon carbide protective housing 704, which is used to improve the high-temperature resistance of the sensor and block the corrosion of coke gas in the furnace.

[0071] The external energy supply and signal processing module 701 includes an antenna system 7011, a signal processing module 7012, a signal transmitting module 7013, and a thermoelectric energy supply module 7014. The thermoelectric energy supply module 7014 is a high-performance thermoelectric material;

[0072] The silicon carbide protective housing 704 is cylindrical. The silicon carbide protective housing 704 includes a first end and a second end. The antenna system 7011 is connected to the first end of the silicon carbide protective housing 704. Inside the first end of the silicon carbide protective housing 704, there is a signal processing module 7012. The front-end probe sensor module 703 is embedded inside the second end of the silicon carbide protective housing 704. In the direction from the outside to the inside, a strip-shaped thermoelectric energy supply module 7014, a cylindrical waveguide channel 702, and a signal transmitting module 7013 are arranged from the silicon carbide protective housing 704 inward. The thermoelectric energy supply module 7014 is connected to the signal processing module 7012. One end of the waveguide channel 702 is connected to the signal processing module 7012, and the other end of the waveguide channel 702 is connected to the front-end probe sensor module 703. The signal transmitting module 7013 is connected to the signal processing module 7012; the antenna system 7011 is used for wireless transmission of signals and systems, power and power transmitters; the signal processing module 7012 is used to process the collected signals and output them through the antenna system 7011; the signal transmitting module 7013 is used to receive the real-time temperature signals collected by the front-end probe sensor module 703 and transmit them to the signal processing module 7012; the thermoelectric energy supply module 7014 uses bismuth telluride Bi2Te3 high-performance thermoelectric material, and the strip-shaped design runs through the high-temperature area and the ambient temperature area of the reserved hole in the furnace, and provides electrical energy for the sensor by means of the temperature difference.

[0073] Combined with Figure 12, an external thread is provided on the outer side of the second end of the silicon carbide protective housing 704. The external thread at the second end of the silicon carbide protective housing 704 meshes with the internal thread of the ceramic fixing nut 102. That is, two split-type wireless passive temperature sensors 7 are fixed on the first plane of the furnace shell, two split-type wireless passive temperature sensors 7 are fixed on the second plane of the furnace shell, and two split-type wireless passive temperature sensors 7 are fixed on the third plane of the furnace shell. There are a total of 6 split-type wireless passive temperature sensors 7 on a single set of furnace cavity furnace shells.

[0074] Step 2: Place a temperature measuring ring at a preset cross-section in each single set of furnace cavities. The preset cross-section can be selected at the center position of the single set of furnace cavities. The intelligent wireless vacuum furnace of the present invention has 8 single sets of furnace cavities, so 8 preset cross-sections are set; operate the intelligent wireless vacuum furnace. During the operation, for each single set of furnace cavities, continuously collect the voltage measurement value of the intelligent wireless vacuum furnace and the current measurement value of the intelligent wireless vacuum furnace through the wireless voltage and current measurement sensors; continuously measure the first temperature of the vacuum degree measuring device through the three-way valve port high-precision wireless temperature sensor, and continuously measure the external furnace vacuum degree measurement value of the intelligent wireless vacuum furnace through the vacuum degree measuring device; continuously collect multiple second temperatures through the split-type wireless passive temperature sensors. The present invention has 6 split-type wireless passive temperature sensors in each single set of furnace cavities, so there are 8 second temperatures; collect the deformation amount of the temperature measuring ring after the operation of the intelligent wireless vacuum furnace ends.

[0075] Step 3: At each running time during the operation, for each single set of furnace cavities, calculate the power based on the voltage measurement value and the current measurement value, and calculate the in-furnace vacuum degree based on the first temperature, multiple second temperatures, and the external furnace vacuum degree measurement value. The power, the in-furnace vacuum degree, and multiple second temperatures form target data. Each single set of furnace cavities corresponds to one target data. Normalize each target data to obtain the normalized target data of each single set of furnace cavities, and further obtain the normalized target data of each single set of furnace cavities corresponding to each running time. Determine the standard temperature through the deformation amount of the temperature measuring ring. Specifically, the deformation amount of the temperature measuring ring and the temperature are in one-to-one correspondence. According to the deformation amount of the temperature measuring ring, its corresponding temperature, that is, the standard temperature, can be queried.

[0076] Among them, calculating the in-furnace vacuum degree based on the first temperature, multiple second temperatures, and the external furnace vacuum degree measurement value includes:

[0077] Step A1: Among multiple second temperatures, determine the second temperatures collected by the two split-type wireless passive temperature sensors fixed on the third plane of the furnace shell, and calculate the average temperature of the second temperatures collected by the two split-type wireless passive temperature sensors on the third plane of the furnace shell.

[0078] Step A2: Calculate the in-furnace vacuum degree based on the average temperature, the first temperature, and the external furnace vacuum degree measurement value, which is specifically realized through the following formula:

[0079]

[0080] Among them, T1` is the first temperature, T2` is the average temperature, T1 and T2 are preset calibration temperatures, P2 is the measurement of the external furnace vacuum degree, and P2` is the internal furnace vacuum degree.

[0081] Step 4: Construct a three-dimensional model of the intelligent wireless vacuum furnace. Based on the normalized target data of each single furnace cavity corresponding to each running time and the three-dimensional model of the intelligent wireless vacuum furnace, perform simulation through simulation software to obtain the temperature distribution data of each preset cross-section corresponding to each running time. For each preset cross-section, correct the temperature distribution data based on the standard temperature and multiple second temperatures to obtain the corrected temperature distribution data, and further obtain the corrected temperature distribution data of each preset cross-section. When the present invention performs correction, it adjusts the temperature in the temperature distribution data to make it close to the standard temperature and the second temperature at the corresponding position, and normalize the corrected temperature distribution data of each preset cross-section to obtain the normalized corrected temperature distribution data of each preset cross-section;

[0082] In the specific implementation process, a three-dimensional model of the intelligent wireless vacuum furnace can be drawn by means of three-dimensional mapping software for the workpiece size and material under the common furnace loading method. Input the normalized target data of each single furnace cavity corresponding to each running time and the three-dimensional model of the intelligent wireless vacuum furnace into the ANSYS simulation software to obtain the temperature distribution data of each preset cross-section corresponding to each running time.

[0083] Step 5: For each running time, use the running time and the normalized target data of each single furnace cavity corresponding to the running time as input samples, and use the normalized corrected temperature distribution data of each preset cross-section corresponding to the running time as output samples. The input samples and the output samples form training samples, and each running time corresponds to one training sample, and then multiple training samples are obtained. The multiple training samples form a data set, and the data set is divided into a training set and a test set according to a preset ratio. Specifically, 80% of the data set can be divided into the training set, and 20% of the data set can be divided into the test set. The training set is used to train the CNN-LSTM prediction model, and the test set is used to evaluate the CNN-LSTM prediction model during the training process.

[0084] The convolutional neural network (CNN) can effectively extract the features of high-dimensional data. After reducing the number of parameters by using convolutional layers and pooling layers, the CNN is input into a long short-term memory network (LSTM) model for prediction. The CNN-LSTM model can improve the computational efficiency and increase the generalization ability. Therefore, the present invention selects the CNN-LSTM prediction model to predict the temperature distribution data. The specific steps are as follows: First, the convolutional neural network (CNN) is used to process the spatial correlation features of the data, and the spatial features are extracted from the input data for downsampling to shrink the sequence, and a temporal feature vector of the high-dimensional mapping space is constructed. Then, the feature sequence extracted by the CNN is input into the LSTM to capture the temporal features of the time series data. The LSTM can model the time-dependent relationship of these feature sequences, so as to better understand the time information in the data. Next, the outputs are merged to form a comprehensive feature representation, and this feature representation is processed through a fully connected layer. Finally, the activation function is used to output continuous predicted values.

[0085] Step 6: Based on multiple training samples in the training set, train the CNN-LSTM prediction model to obtain the trained CNN-LSTM prediction model;

[0086] Input the input samples in the training samples into the CNN-LSTM prediction model to obtain the predicted temperature distribution data. According to the predicted temperature distribution data and the corrected temperature distribution data, calculate the loss function, and determine whether the loss function is greater than the preset threshold. In the case where the loss function is not greater than the preset threshold, obtain the trained CNN-LSTM prediction model. In the case where the loss function is greater than the preset threshold, update the parameters of the CNN-LSTM prediction model based on the loss function, and obtain the next training sample, and return to execute: Input the input samples in the training samples into the CNN-LSTM prediction model.

[0087] Step 7: For different workpiece sizes and materials of the intelligent wireless vacuum furnace, execute Steps 2-6 to obtain multiple trained CNN-LSTM prediction models under different workpiece sizes and materials;

[0088] Step 8: Combine Figure 16 , run the intelligent wireless vacuum furnace to be predicted, and obtain the normalized data to be predicted for each single furnace cavity corresponding to each running time through the wireless voltage and current measurement sensors, three-way valve port high-precision wireless temperature sensors, vacuum degree measurement device, and split-type wireless passive temperature sensors of the intelligent wireless vacuum furnace;

[0089] Continuously collect the target voltage value of the intelligent wireless vacuum furnace to be predicted and the target current value of the intelligent wireless vacuum furnace to be predicted through a wireless voltage and current measurement sensor; continuously measure the first target temperature of the vacuum degree measurement device through a three-way valve port high-precision wireless temperature sensor, and continuously measure the target external vacuum degree of the intelligent wireless vacuum furnace through the vacuum degree measurement device; continuously collect multiple second target temperatures through a split-type wireless passive temperature sensor.

[0090] At each running time during the operation process, for each single furnace cavity, calculate the target power based on the target voltage value and the target current value, and calculate the target internal vacuum degree based on the first target temperature, multiple second target temperatures, and the target external vacuum degree. The target power, the target external vacuum degree, and the multiple second target temperatures form the data to be predicted. Normalize each data to be predicted. Each single furnace cavity corresponds to a data to be predicted, and the normalized data to be predicted is obtained, and then the normalized data to be predicted corresponding to each single furnace cavity at each running time is obtained.

[0091] Step 9: According to the workpiece size and material of the intelligent wireless vacuum furnace to be predicted, obtain the trained CNN-LSTM prediction model corresponding to the workpiece size and material. Input the running time and the normalized data to be predicted of each single furnace cavity at this running time into the obtained trained CNN-LSTM prediction model, and output the temperature distribution data of the preset cross-section in each single furnace cavity of the intelligent wireless vacuum furnace to be predicted. Specifically, according to the preset quantity, 36-point data of each cross-section can be output. Denormalize the temperature distribution data of each preset cross-section to obtain the denormalized temperature distribution data of each preset cross-section, as Figure 17 shown in the cross-section temperature point distribution diagram of the vacuum sintering furnace.

[0092] Step 10: Visualize and output the denormalized temperature distribution data of each preset cross-section.

[0093] Specifically, with the help of LabVIEW, convert the cross-section temperature data into a real-time cloud map for display. To improve the resolution when converting each temperature cross-section into a cloud map, select the Kriging spatial interpolation method to dynamically adjust the resolution of the temperature cloud map, which can reduce the amount of data required for prediction and reduce the prediction time; calculate the non-uniformity of the temperature data within each cross-section and feedback the result to the PID, and adjust the power of the heating element group corresponding to the cross-section through the PID to ensure the uniformity of the temperature in the working area.

[0094] Combined with Figure 18, the visualization platform includes a real-time cloud map of the temperature field of the eight-zone cross-section, prediction start and stop buttons, buttons for switching prediction models of different charging methods, and buttons for switching the resolution of the real-time cloud map. The visualization platform is used for real-time display of the temperatures of 36 points at the central cross-section of eight independent zones in the furnace; switching the charging method to select prediction models for different workpiece sizes and materials; switching the resolution to dynamically adjust the resolution of the temperature cloud map by the Kriging spatial interpolation method (interpolation from 36 points to more than 1000 points).

[0095] The above description is only a preferred embodiment of the present disclosure and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the technical solutions formed by mutually replacing the above features with the (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A temperature prediction and visualization method for an intelligent wireless vacuum furnace, characterized in that, Including: Step 1: Construct an intelligent wireless vacuum furnace, which includes 8 single-group furnace cavities. Each single-group furnace cavity includes a furnace shell, carbon felt, graphite plates, a heating element group, a power connection device, a vacuum degree measuring device, and multiple split-type wireless passive temperature sensors. Among them, the power connection device includes wireless voltage and current measuring sensors, and the vacuum degree measuring device includes a high-precision wireless temperature sensor at the three-way valve port and a vacuum gauge measuring sensor; Step 2: Place temperature measurement rings at preset cross-sections in each single-group furnace cavity. Run the intelligent wireless vacuum furnace. During the operation, for each single-group furnace cavity, continuously collect the voltage measurement value of the intelligent wireless vacuum furnace and the current measurement value of the intelligent wireless vacuum furnace through the wireless voltage and current measuring sensors; continuously measure the first temperature of the vacuum degree measuring device through the high-precision wireless temperature sensor at the three-way valve port, and continuously measure the furnace external vacuum degree measurement value of the intelligent wireless vacuum furnace through the vacuum degree measuring device; continuously collect multiple second temperatures through the split-type wireless passive temperature sensors; collect the deformation amount of the temperature measurement ring after the operation of the intelligent wireless vacuum furnace ends; Step 3: At each running time during the operation, for each single-group furnace cavity, calculate the power based on the voltage measurement value and the current measurement value, and calculate the in-furnace vacuum degree based on the first temperature, multiple second temperatures, and the furnace external vacuum degree measurement value. The power, in-furnace vacuum degree, and multiple second temperatures form target data. Each single-group furnace cavity corresponds to one target data. Normalize each target data to obtain the normalized target data of each single-group furnace cavity, and then obtain the normalized target data of each single-group furnace cavity corresponding to each running time. Determine the standard temperature through the deformation amount of the temperature measurement ring; Step 4: Construct a three-dimensional model of the intelligent wireless vacuum furnace. Based on the normalized target data of each single-group furnace cavity corresponding to each running time and the three-dimensional model of the intelligent wireless vacuum furnace, perform simulation through simulation software to obtain the temperature distribution data of each preset cross-section corresponding to each running time. For each preset cross-section, correct the temperature distribution data based on the standard temperature and multiple second temperatures to obtain the corrected temperature distribution data, and then obtain the corrected temperature distribution data of each preset cross-section. Normalize the corrected temperature distribution data of each preset cross-section to obtain the normalized corrected temperature distribution data of each preset cross-section; Step 5: For each running time, use the running time and the normalized target data of each single-group furnace cavity corresponding to this running time as input samples, and use the normalized corrected temperature distribution data of each preset cross-section corresponding to this running time as output samples. The input samples and the output samples form training samples. Each running time corresponds to one training sample, and then obtain multiple training samples. The multiple training samples form a data set. Divide the data set into a training set and a test set according to a preset ratio; Step 6: Train the CNN-LSTM prediction model based on multiple training samples in the training set to obtain the trained CNN-LSTM prediction model; Step 7: For different workpiece sizes and materials of the intelligent wireless vacuum furnace, execute Steps 2-6 to obtain multiple trained CNN-LSTM prediction models under different workpiece sizes and materials; Step 8: Operate the intelligent wireless vacuum furnace to be predicted. Through the wireless voltage and current measurement sensors, high-precision wireless temperature sensors at the three-way valve ports, vacuum measurement devices, and split-type wireless passive temperature sensors of the intelligent wireless vacuum furnace, obtain the normalized data to be predicted for each single furnace cavity corresponding to each running time; Step 9: According to the workpiece size and material of the intelligent wireless vacuum furnace to be predicted, obtain the trained CNN-LSTM prediction model corresponding to the workpiece size and material. Input the running time and the normalized data to be predicted for each single furnace cavity at this running time into the obtained trained CNN-LSTM prediction model, output the temperature distribution data of the preset cross-section in each single furnace cavity of the intelligent wireless vacuum furnace to be predicted, and perform denormalization on the temperature distribution data of each preset cross-section to obtain the denormalized temperature distribution data of each preset cross-section; Step 10: Visually output the denormalized temperature distribution data of each preset cross-section.

2. The temperature prediction and visualization method of an intelligent wireless vacuum furnace according to claim 1, wherein The furnace shell of the single furnace cavity is connected to the power connection device. In the order from outside to inside, the single furnace cavity from the furnace shell inward is carbon felt, graphite plate, and heating element group; the furnace shell of the single furnace cavity is provided with wireless passive temperature sensor mounting holes, and the wireless passive temperature sensor mounting holes include left mounting holes, top mounting holes, and right mounting holes. The furnace shell of the single furnace cavity includes four planes, namely the first plane, the second plane, the third plane, and the fourth plane. The first plane and the third plane are parallel, the second plane and the fourth plane are parallel, the second plane is perpendicular to both the first plane and the third plane. The first plane is provided with two left mounting holes, the second plane is provided with two top mounting holes, the third plane is provided with two right mounting holes, ceramic fixing nuts are fixed above the left mounting holes, top mounting holes, and right mounting holes, and the third plane is provided with a vacuum gauge mounting hole.

3. The temperature prediction and visualization method of an intelligent wireless vacuum furnace according to claim 2, wherein The vacuum measurement device includes a detachable vacuum gauge port pipe, a three-layer honeycomb filter molybdenum plate, an in-furnace connecting molybdenum pipe, an out-of-furnace water-cooled pipe, a three-way valve, a high-precision wireless temperature sensor at the three-way valve port, and a vacuum gauge measurement sensor; the three-layer honeycomb filter molybdenum plate includes a first-layer honeycomb filter molybdenum plate, a second-layer honeycomb filter molybdenum plate, and a third-layer honeycomb filter molybdenum plate. Among them, the pore size of the internal pores in each layer of the honeycomb filter molybdenum plate is the same, and the pore size of the internal pores in the first-layer honeycomb filter molybdenum plate, the pore size of the internal pores in the second-layer honeycomb filter molybdenum plate, and the pore size of the internal pores in the third-layer honeycomb filter molybdenum plate are all different; The out-of-furnace water-cooled pipe is provided with a water-cooled pipe water inlet and a water-cooled pipe water outlet. Flange plates are provided at both ends of the out-of-furnace water-cooled pipe. The three-way valve is provided with three ports, namely the first port, the second port, and the third port. Flange plates are provided at the first port, the second port, and the third port; The detachable vacuum gauge port tube is fixed at the vacuum gauge mounting hole through a flange; the three-layer honeycomb filter molybdenum plate is embedded in the detachable vacuum gauge port tube, the detachable vacuum gauge port tube is embedded in the first end of the in-furnace connecting molybdenum tube, a flange is provided at the second end of the in-furnace connecting molybdenum tube, the flange at the second end of the in-furnace connecting molybdenum tube is connected to the flange at one end of the out-of-furnace water-cooled pipe, the flange at the other end of the out-of-furnace water-cooled pipe is connected to the flange at the first port of the three-way valve, a flange is provided at the port of the three-way valve with high-precision wireless temperature sensing, the flange at the second port of the three-way valve is connected to the flange at the port of the three-way valve with high-precision wireless temperature sensing, a flange is provided at the port of the vacuum gauge measurement sensor, and the flange at the third port of the three-way valve is connected to the flange at the port of the vacuum gauge measurement sensor.

4. A temperature prediction and visualization method for an intelligent wireless vacuum furnace according to claim 2, characterized in that The split-type wireless passive temperature sensor includes an external energy supply and signal processing module, a waveguide channel, a front-end probe sensor module, and a silicon carbide protective housing; The front-end probe sensor module includes an alumina protective film, a diamond film, interdigital electrodes, and a sapphire substrate. The interdigital electrodes are placed on the sapphire substrate, the diamond film is deposited on the interdigital electrodes and the sapphire substrate, and the alumina protective film is deposited on the diamond film; The external energy supply and signal processing module includes an antenna system, a signal processing module, a signal transmission module, and a thermoelectric energy supply module. The thermoelectric energy supply module is a high-performance thermoelectric material; The silicon carbide protective housing is cylindrical and includes a first end and a second end. The antenna system is connected to the first end of the silicon carbide protective housing. A signal processing module is provided inside the first end of the silicon carbide protective housing. The front-end probe sensor module is embedded inside the second end of the silicon carbide protective housing. In the direction from the outside to the inside, a strip-shaped thermoelectric energy supply module, a cylindrical waveguide channel, and a signal transmission module are arranged from the silicon carbide protective housing inward. The thermoelectric energy supply module is connected to the signal processing module. One end of the waveguide channel is connected to the signal processing module, the other end of the waveguide channel is connected to the front-end probe sensor module, and the signal transmission module is connected to the signal processing module; External threads are provided on the outer side of the second end of the silicon carbide protective housing, and the external threads at the second end of the silicon carbide protective housing are engaged with the internal threads of the ceramic fixing nut.

5. The temperature prediction and visualization method of an intelligent wireless vacuum furnace according to claim 4, characterized in that In step 3, based on the first temperature, multiple second temperatures, and the measured value of the out-of-furnace vacuum degree, the in-furnace vacuum degree is calculated, including: Step A1: Among the multiple second temperatures, determine the second temperatures collected by the two split-type wireless passive temperature sensors fixed on the third plane of the furnace shell, and calculate the average temperature of the second temperatures collected by the two split-type wireless passive temperature sensors on the third plane of the furnace shell; Step A2: Based on the average temperature, the first temperature, and the measured value of the out-of-furnace vacuum degree, the in-furnace vacuum degree is calculated, which is specifically realized through the following formula: Where, T1` is the first temperature, T2` is the average temperature, T1 and T2 are preset calibration temperatures, P2 is the measured out-of-furnace vacuum degree, and P2` is the in-furnace vacuum degree.

6. A temperature prediction and visualization method for an intelligent wireless vacuum furnace according to claim 1, characterized in that Step 6 specifically includes: Input the input samples in the training samples into the CNN-LSTM prediction model to obtain predicted temperature distribution data. Calculate the loss function based on the predicted temperature distribution data and the corrected temperature distribution data. Determine whether the loss function is greater than a preset threshold. If the loss function is not greater than the preset threshold, obtain the trained CNN-LSTM prediction model. If the loss function is greater than the preset threshold, update the parameters of the CNN-LSTM prediction model based on the loss function, obtain the next training sample, and return to execute: input the input samples in the training samples into the CNN-LSTM prediction model.

7. A temperature prediction and visualization method for an intelligent wireless vacuum furnace according to claim 1, characterized in that Step 8 specifically includes: Continuously collect the target voltage value of the intelligent wireless vacuum furnace to be predicted and the target current value of the intelligent wireless vacuum furnace to be predicted through a wireless voltage and current measurement sensor; continuously measure the first target temperature of the vacuum degree measurement device through a three-way valve port high-precision wireless temperature sensor, and continuously measure the target external vacuum degree of the intelligent wireless vacuum furnace through the vacuum degree measurement device; continuously collect multiple second target temperatures through a split-type wireless passive temperature sensor; At each running time during the operation, for each single furnace cavity, calculate the target power based on the target voltage value and the target current value, calculate the target internal vacuum degree based on the first target temperature, multiple second target temperatures, and the target external vacuum degree. The target power, the target external vacuum degree, and the multiple second target temperatures form the data to be predicted. Normalize each data to be predicted. Each single furnace cavity corresponds to one data to be predicted to obtain the normalized data to be predicted, and further obtain the normalized data to be predicted for each single furnace cavity corresponding to each running time.

Images