Automatic preparation method and system for integrated temperature and liquid level detection probe

Abstract

The invention relates to the technical field of detection probe preparation, in particular to an automatic preparation method and system for an integrated temperature and liquid level detection probe, and the method comprises the steps: carrying out the integration operation on a plurality of sensor groups and a printed circuit board, carrying out the adhesive coating operation on an integrated circuit board, and carrying out the board separation operation on the adhesive circuit board based on the plurality of sensor groups. A target detection assembly is aligned with a target probe shell and pressurization operation is executed, heating and curing operation is executed on an assembly shell, welding operation is executed on a curing detection assembly and a target signal cable, and pouring sealant pouring operation is executed on a signal detection assembly; and performing quality control screening operation based on detection performance test and packaging inspection on the plurality of initial detection probes to obtain a plurality of temperature and liquid level detection probes. According to the integrated temperature and liquid level detection probe, the gold wire bonding process parameters are predicted through analogue simulation, and the process parameters are corrected when the expected performance is not achieved, so that the integrated temperature and liquid level detection probe with better performance and better stability is obtained.

Description

Technical Field

[0001] This invention relates to the field of detection probe fabrication technology, and in particular to an automated fabrication method and system for an integrated temperature and liquid level detection probe. Background Technology

[0002] Temperature and liquid level detection probes are widely used as sensor devices in the field of home appliances. Their performance stability and manufacturing precision are directly related to the accuracy and reliability of the detection data.

[0003] In the existing technology, temperature and liquid level detection probes are prepared by dispersing temperature sensors and liquid level sensors on the probe housing and manually welding the positive and negative electrodes together.

[0004] While the aforementioned methods can fabricate temperature and level detection probes, they have significant limitations: low integration of the probe's internal structure, insufficient automation in the fabrication process, heavy reliance on manual labor, and a lack of systematic control and optimization of gold wire bonding process parameters. This results in low product consistency and production efficiency, making it difficult to fundamentally guarantee product yield and long-term reliability. Therefore, there is a need for an integrated temperature and level detection probe that can predict gold wire bonding process parameters through simulation and correct these parameters when expected performance is not achieved, thereby obtaining a probe with superior performance and better stability. Summary of the Invention

[0005] This invention provides an automated manufacturing method for an integrated temperature and liquid level detection probe and a computer-readable storage medium. Its main purpose is to predict the gold wire bonding process parameters through simulation and to correct the process parameters when the expected performance is not achieved, thereby obtaining an integrated temperature and liquid level detection probe with better performance and stability.

[0006] To achieve the above objectives, the present invention provides an automated manufacturing method for an integrated temperature and liquid level detection probe, comprising:

[0007] The system receives a probe preparation instruction and confirms the preparation environment based on the instruction. The preparation environment includes multiple sensor groups, printed circuit boards, multiple probe housings, and multiple signal cables. The sensor groups include a temperature sensor and a liquid level sensor.

[0008] An integration operation is performed on multiple sensor groups and printed circuit boards to obtain an integrated circuit board;

[0009] An adhesive application operation is performed on the integrated circuit board to obtain an adhesive circuit board. Based on multiple sensor groups, a board separation operation is performed on the adhesive circuit board to obtain multiple detection components, wherein the sensor groups and detection components correspond one-to-one.

[0010] Extract one detection component from multiple detection components sequentially to obtain the target detection component, and perform the following operations on the target detection component:

[0011] Extract the probe housing corresponding to the detection component from multiple probe housings to obtain the target probe housing;

[0012] Align the target detection component with the target probe housing and apply pressure to obtain the component housing. Then, perform a heat curing operation on the component housing to obtain the cured detection component.

[0013] The signal cable corresponding to the curing detection component is identified from multiple signal cables to obtain the target signal cable. A soldering operation is then performed on the curing detection component and the target signal cable to obtain the signal detection component.

[0014] A potting compound is applied to the signal detection component to obtain a potted detection probe. A heat curing operation is then performed on the potted detection probe to obtain an initial detection probe.

[0015] The initial detection probes are summarized to obtain multiple initial detection probes;

[0016] Multiple initial detection probes were subjected to quality control screening based on detection performance testing and packaging inspection to obtain multiple temperature and liquid level detection probes.

[0017] Optionally, the integration operation of multiple sensor groups and printed circuit boards to obtain an integrated circuit board includes:

[0018] Identify the integrated environment, which includes the printer, pick-and-place machine, and reflow oven;

[0019] Solder paste is applied to the printed circuit board using a printing press to obtain a solder paste circuit board. The solder paste circuit board is then fed into a pick-and-place machine, which performs an image acquisition operation on the solder paste circuit board to obtain a circuit board image.

[0020] A contour extraction operation is performed on the circuit board image to obtain multiple contours to be confirmed. A comparison operation is performed on the multiple contours to be confirmed using the pre-constructed temperature pad contour and the pre-constructed liquid level pad contour to obtain multiple target temperature pad contours and multiple target liquid level pad contours.

[0021] Based on the pre-constructed solder paste circuit board coordinate system, multiple target temperature pad contours and multiple target liquid level pad contours, multiple pad position groups are determined in the solder paste circuit board. Each pad position group includes a temperature pad position and a liquid level pad position. The temperature pad position corresponds one-to-one with the target temperature pad contour, and the liquid level pad position corresponds one-to-one with the target liquid level pad.

[0022] Extract one sensor group from multiple sensor groups sequentially to obtain the target sensor group, and perform the following operations on the target sensor group:

[0023] The target pad position group is obtained by determining the pad position group corresponding to the target sensor group from multiple pad position groups;

[0024] A pick-and-place machine is used to pick up the target sensor group, and an alignment and placement operation is performed on the picked-up target sensor group and the target pad position group to obtain the sensor pads;

[0025] Summarize the sensor pads to obtain multiple sensor pads, and denote these multiple sensor pads as the sensor circuit board.

[0026] The sensor circuit board was cured using a reflow oven to obtain a cured circuit board;

[0027] Obtain the gold wire bonding process parameter set, perform gold wire bonding operation based on the gold wire bonding process parameter set and the solidified circuit board, and obtain the integrated circuit board.

[0028] Optionally, obtaining the gold wire bonding process parameter set includes:

[0029] A knowledge graph of gold wire bonding is constructed based on pre-built historical data of gold wire bonding;

[0030] Extract multiple process parameters, their value ranges, and constraint rules from the knowledge graph;

[0031] Extract a process parameter value range from multiple process parameter value ranges sequentially to obtain the target value range, and then perform the following operations on the target value range:

[0032] Based on the preset number of process parameters, the target value range is divided into intervals to obtain multiple value intervals. One value interval is extracted from the multiple value intervals in turn, and a numerical extraction operation is performed on the extracted value interval to obtain the process parameter values. The process parameter values ​​are summarized to obtain multiple process parameter values ​​corresponding to the target value range. The multiple process parameter values ​​are recorded as the target process scattered value group.

[0033] Summarize the target process scattered value groups to obtain multiple target process scattered value groups corresponding to multiple process parameter value ranges. Perform parameter classification operation based on element-distinctness on multiple target process scattered value groups based on multiple process parameters to obtain multiple initial process parameter sets. The initial process parameter sets include multiple initial process parameters, and the initial process parameters correspond one-to-one with the process parameters.

[0034] If multiple initial process parameter sets all meet the constraint rules, then the multiple initial process parameter sets are confirmed as multiple process parameter sets.

[0035] The gold wire bonding process parameter set is obtained based on multiple process parameter sets.

[0036] Optionally, obtaining the gold wire bonding process parameter set based on multiple process parameter sets includes:

[0037] Multiple candidate process parameter sets are generated based on multiple process parameter sets;

[0038] Simulation operations are performed on multiple candidate process parameter sets to obtain multiple simulated integrated circuit boards, wherein the simulated integrated circuit boards and candidate process parameter sets correspond one-to-one;

[0039] A board splitting operation is performed on multiple simulated integrated circuit boards to obtain multiple simulated testing component sets. Each simulated testing component set includes multiple simulated testing components, and each simulated testing component set corresponds one-to-one with a simulated integrated circuit board.

[0040] Performance prediction is performed on multiple simulation detection component sets to obtain multiple performance prediction sets. Each performance prediction set includes multiple performance prediction groups. There is a one-to-one correspondence between the simulation detection component sets and the performance prediction sets. Each performance prediction group includes multiple performance prediction values, and there is a one-to-one correspondence between the performance prediction groups and the simulation detection components.

[0041] Multiple performance qualities are calculated based on multiple performance prediction sets and pre-built performance quality calculation formulas. The multiple performance qualities are compared with the pre-built performance quality threshold. If there is a performance quality among the multiple performance qualities that is greater than or equal to the performance quality threshold, the performance quality with the largest value among the multiple performance qualities is extracted to obtain the maximum performance quality.

[0042] The candidate process parameter set corresponding to the maximum performance quality is determined from the multiple candidate process parameter sets to obtain the gold wire bonding process parameter set;

[0043] If multiple performance quality values ​​are all less than the performance quality threshold, then multiple candidate process parameter sets are used as the multiple process parameter sets, and the step of generating multiple candidate process parameter sets based on multiple process parameter sets is returned until the gold wire bonding process parameter set is obtained.

[0044] Optionally, the calculation of multiple performance qualities based on multiple performance prediction sets and pre-built performance quality calculation formulas includes:

[0045] Extract a performance prediction set from multiple performance prediction sets sequentially to obtain the target performance prediction set, and then perform the following operations on the target performance prediction set:

[0046] For each performance prediction group in the target performance prediction group set, a vector construction operation is performed to obtain multiple prediction performance vectors, and a prediction performance matrix is ​​generated based on the multiple prediction performance vectors.

[0047] Multiple performance types are identified based on the predicted performance matrix, multiple weights are assigned to the multiple performance types, and a weight vector is constructed using the multiple weights.

[0048] Calculate the weight performance matrix based on the weight vector and the prediction performance matrix;

[0049] Normalize the weight performance matrix to obtain the normalized performance matrix, as shown below:

[0050] ,

[0051] in, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the performance prediction group An index of normalized performance values, Indicates the first in the target performance prediction group set Indexes for performance prediction groups;

[0052] By summing the normalized performance matrices, multiple normalized performance matrices are obtained;

[0053] Multiple performance qualities are calculated based on multiple normalized performance matrices and performance quality calculation formulas.

[0054] Optionally, the calculation of multiple performance qualities based on multiple normalized performance matrices and performance quality calculation formulas includes:

[0055] Extract one of the multiple performance types sequentially to obtain the target performance type, and perform the following operations on the target performance type:

[0056] Extract multiple normalized performance values ​​corresponding to the target performance type from the multiple normalized performance matrices, and identify the normalized performance value with the largest value and the normalized performance value with the smallest value among the multiple normalized performance values ​​to obtain the maximum performance value and the minimum performance value.

[0057] The maximum and minimum performance values ​​are summarized separately to obtain the maximum performance value vector and minimum performance value vector corresponding to multiple performance types. The maximum performance value vector includes multiple maximum value elements, and each maximum value element corresponds one-to-one with the performance type.

[0058] Extract one of the multiple normalized performance matrices sequentially to obtain the target normalized performance matrix. Perform the following operations on the target normalized performance matrix:

[0059] Based on the multiple simulation detection component sets, multiple normalized performance vectors are obtained from the target normalized performance matrix. Then, one normalized performance vector is extracted from the multiple normalized performance vectors to obtain the target performance vector. The following operations are performed on the target performance vector:

[0060] Calculate the Euclidean distance between the target performance vector and the maximum performance value vector to obtain the positive ideal distance; calculate the Euclidean distance between the target performance vector and the minimum performance value vector to obtain the negative ideal distance.

[0061] Performance similarity is calculated based on positive and negative ideal distances;

[0062] Summarize the performance similarity scores to obtain multiple performance similarity scores, and construct the performance similarity vector corresponding to the target normalized performance matrix based on the performance similarity scores;

[0063] By summing the performance proximity vectors, multiple performance proximity vectors corresponding to multiple performance prediction sets are obtained, where there is a one-to-one correspondence between the performance prediction sets and the performance proximity vectors.

[0064] Multiple performance qualities are calculated based on multiple performance proximity vectors and performance quality calculation formulas.

[0065] Optionally, the calculation of multiple performance qualities based on multiple performance proximity vectors and performance quality calculation formulas includes:

[0066] Extract one of the multiple normalized performance matrices sequentially to obtain the key normalized performance matrix. Perform the following operations on the key normalized performance matrix:

[0067] Based on the multiple performance types, multiple key performance vectors are obtained from the key normalized performance matrix. Each of the multiple key performance vectors is the transpose of the column vector of the key normalized performance matrix, and the key performance vectors correspond one-to-one with the performance types. Each key performance vector includes multiple key performance values.

[0068] Extract one key performance vector from multiple key performance vectors sequentially to obtain the target key performance vector, and then perform the following operations on the target key performance vector:

[0069] Calculate the standard deviation of multiple key performance values ​​to obtain the performance standard deviation;

[0070] Summarize the performance standard deviations to obtain multiple performance standard deviations corresponding to multiple performance types, and construct a performance standard deviation vector based on the multiple performance standard deviations;

[0071] The performance proximity vector corresponding to the normalized performance matrix represented by the performance standard deviation vector is extracted from multiple performance proximity vectors to obtain a definite proximity vector. The performance quality is then calculated based on the performance standard deviation vector, the definite proximity vector, and the performance quality calculation formula.

[0072] By summarizing the performance qualities, multiple performance qualities are obtained, each of which corresponds one-to-one with a set of candidate process parameters.

[0073] Optionally, the performance quality calculation formula is as follows:

[0074] ,

[0075] in, Indicates performance quality. Indicates the weight of performance similarity. Indicates the standard deviation weight of performance. Indices representing multiple performance similarity values ​​in the performance similarity vector. This represents the first of several proximity values ​​in the performance proximity vector. A proximity index, The first element in the performance proximity vector represents the... Performance similarity Indices representing multiple performance standard deviations of the performance standard deviation vector. The first of several performance standard deviations in the performance standard deviation vector represents the first of the several performance standard deviations. An index with a performance standard deviation, The first element in the performance standard deviation vector represents the... One performance standard deviation.

[0076] Optionally, the step of performing gold wire bonding operation based on the gold wire bonding process parameter set and the cured circuit board to obtain an integrated circuit board includes:

[0077] The gold wire bonding process parameter set is imported into the pre-built gold wire bonding machine to obtain the setting bonding machine. The setting bonding machine is used to confirm the positions of multiple sensor groups and multiple bonding position groups in the cured circuit board. Among them, the bonding position group includes a temperature bonding position and a liquid level bonding position.

[0078] The location of a sensor group is determined sequentially from multiple sensor group locations to obtain the target sensor group location. The following operations are then performed on the target sensor group location:

[0079] Temperature metal wire and liquid level metal wire are obtained, and a discharge operation is performed on the temperature metal wire and liquid level metal wire using a setting bonding machine to obtain temperature metal ball and liquid level metal ball.

[0080] The temperature metal ball and the liquid level metal ball are pressed together to the target sensor group position to obtain the temperature welding metal wire and the liquid level welding metal wire;

[0081] Determine the target binding position group corresponding to the target sensor group position among multiple binding position groups to obtain the target binding position group;

[0082] The temperature welding metal wire and the liquid level welding metal wire are welded to the target bonding position group to obtain the gold wire bonding pad.

[0083] The gold wire bonding pads are aggregated to obtain multiple gold wire bonding pads, which are then referred to as the integrated circuit board.

[0084] To achieve the above objectives, the present invention also provides an automated manufacturing system for an integrated temperature and liquid level detection probe, comprising:

[0085] The instruction receiving module is used to receive the detection probe preparation instruction and confirm the preparation environment based on the detection probe preparation instruction. The preparation environment includes multiple sensor groups, printed circuit boards, multiple probe housings and multiple signal cables. The sensor group includes a temperature sensor and a liquid level sensor.

[0086] The integrated board separation module is used to perform integration operations on multiple sensor groups and printed circuit boards to obtain an integrated circuit board. Adhesive is applied to the integrated circuit board to obtain an adhesive circuit board. Based on multiple sensor groups, the adhesive circuit board is separated to obtain multiple detection components. The sensor groups and detection components correspond one-to-one.

[0087] The encapsulation module is used to sequentially extract one detection component from multiple detection components to obtain a target detection component. The following operations are performed on the target detection component: extracting the probe housing corresponding to the detection component from multiple probe housings to obtain a target probe housing; aligning the target detection component with the target probe housing and applying pressure to obtain a component housing; performing a heat curing operation on the component housing to obtain a cured detection component; identifying the signal cable corresponding to the cured detection component from multiple signal cables to obtain a target signal cable; performing a soldering operation on the cured detection component and the target signal cable to obtain a signal detection component; performing a potting compound injection operation on the signal detection component to obtain a potted detection probe; performing a heat curing operation on the potted detection probe to obtain an initial detection probe; and summing the initial detection probes to obtain multiple initial detection probes.

[0088] The quality inspection module is used to perform quality control screening operations based on detection performance testing and packaging inspection on multiple initial detection probes to obtain multiple temperature and liquid level detection probes.

[0089] To address the problems described in the background art, multiple sensor groups and printed circuit boards are integrated to obtain an integrated circuit board. It is evident that this invention integrates multiple sensor groups onto a single printed circuit board, solving the problem of low integration density within the probe's internal structure. Simultaneously, this invention replaces the traditional error-prone manual soldering by accurately predicting the gold wire bonding process parameters, fundamentally ensuring high consistency in the electrical performance of each temperature and level detection probe. Furthermore, when the performance of the temperature and level detection probe fails to meet expectations, the gold wire bonding process parameters are corrected, resulting in an integrated temperature and level detection probe with superior performance and better stability. This invention integrates multiple traditionally serial and separate bonding and curing steps into a highly automated surface mount and gold wire bonding process, significantly improving production efficiency. Therefore, this invention predicts the gold wire bonding process parameters through simulation and corrects the parameters when expected performance is not achieved, thereby obtaining an integrated temperature and level detection probe with superior performance and better stability. Attached Figure Description

[0090] Figure 1 A schematic flowchart illustrating an automated manufacturing method for an integrated temperature and liquid level detection probe according to an embodiment of the present invention;

[0091] Figure 2 This is a schematic diagram of the external structure of an integrated temperature and liquid level detection probe provided in an embodiment of the present invention;

[0092] Figure 3 This is a circuit diagram of an integrated temperature and liquid level detection probe provided according to an embodiment of the present invention.

[0093] Explanation of reference numerals in the attached figures:

[0094] 7. Probe housing; 11. First housing; 12. Second housing; 23. Printed circuit board; 31. Adhesive fixing layer; 32. Buffer protective layer; 41. Signal negative cable.

[0095] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0096] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0097] This application provides an automated manufacturing method for an integrated temperature and liquid level detection probe. The executing entity of this automated manufacturing method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the automated manufacturing method for the integrated temperature and liquid level detection probe can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0098] Reference Figure 1 The diagram shown is a flowchart illustrating an automated fabrication method for an integrated temperature and liquid level detection probe according to an embodiment of the present invention. In this embodiment, the automated fabrication method for the integrated temperature and liquid level detection probe includes:

[0099] S1. Receive the probe preparation instruction and confirm the preparation environment based on the probe preparation instruction. The preparation environment includes multiple sensor groups, printed circuit boards, multiple probe housings and multiple signal cables. The sensor groups include a temperature sensor and a liquid level sensor.

[0100] Specifically, the temperature and liquid level detection probe includes 7. a probe housing; 11. a first housing; 12. a second housing; 23. a printed circuit board; 31. an adhesive fixing layer; 32. a buffer protective layer; and 41. a signal negative cable. For details, please refer to [link / reference needed]. Figure 2 .

[0101] It should be noted that the probe housing is assembled from a first housing and a second housing. The adhesive fixing layer is used to secure the detection component to the probe housing. The buffer protective layer is used to protect the detection component from damage or failure due to significant external forces during use. The signal negative cable is used to connect the detection component and the printed circuit board, enabling signal transmission between the two.

[0102] It's clear that a temperature and level detection probe refers to a probe that can simultaneously detect both temperature and level. The probe housing is a protective enclosure for the detection components. The detection components consist of a level sensor, a temperature sensor, and a printed circuit board. A temperature sensor is used to detect temperature, such as an NTC temperature sensor. A level sensor is used to detect liquid level, such as a piezoelectric ceramic plate. The printed circuit board is the carrier used to connect the temperature sensor and the level sensor. The encapsulation structure is the structure that wraps around the detection components, protecting them. Signal cables are the wires used to transmit current and signals.

[0103] It should be noted that the probe preparation command is a manually initiated command used to automate the preparation of temperature and level detection probes, and the preparation environment refers to the environment used to automate the preparation of temperature and level detection probes.

[0104] Understandably, a sensor group refers to a collection of temperature sensors and level sensors, and adhesive is a glue used to fix the detection components to the probe housing, such as epoxy resin glue.

[0105] For example, employees at the temperature and level detection probe manufacturing plant send probe manufacturing instructions, and the computer receives these instructions. When a certain automated manufacturing step is missing (e.g., insufficient adhesive or insufficient number of sensor groups), the computer confirms the manufacturing environment, thereby ensuring the integrity and continuity of the entire automated manufacturing process for the temperature and level detection probe.

[0106] S2. Perform integration operations on multiple sensor groups and printed circuit boards to obtain an integrated circuit board.

[0107] Understandably, an integrated circuit board is a circuit board that integrates multiple sensor groups onto a printed circuit board.

[0108] Furthermore, the integration operation performed on multiple sensor groups and printed circuit boards to obtain an integrated circuit board includes:

[0109] Identify the integrated environment, which includes the printer, pick-and-place machine, and reflow oven;

[0110] Solder paste is applied to the printed circuit board using a printing press to obtain a solder paste circuit board. The solder paste circuit board is then fed into a pick-and-place machine, which performs an image acquisition operation on the solder paste circuit board to obtain a circuit board image.

[0111] A contour extraction operation is performed on the circuit board image to obtain multiple contours to be confirmed. A comparison operation is performed on the multiple contours to be confirmed using the pre-constructed temperature pad contour and the pre-constructed liquid level pad contour to obtain multiple target temperature pad contours and multiple target liquid level pad contours.

[0112] Based on the pre-constructed solder paste circuit board coordinate system, multiple target temperature pad contours and multiple target liquid level pad contours, multiple pad position groups are determined in the solder paste circuit board. Each pad position group includes a temperature pad position and a liquid level pad position. The temperature pad position corresponds one-to-one with the target temperature pad contour, and the liquid level pad position corresponds one-to-one with the target liquid level pad.

[0113] Extract one sensor group from multiple sensor groups sequentially to obtain the target sensor group, and perform the following operations on the target sensor group:

[0114] The target pad position group is obtained by determining the pad position group corresponding to the target sensor group from multiple pad position groups;

[0115] A pick-and-place machine is used to pick up the target sensor group, and an alignment and placement operation is performed on the picked-up target sensor group and the target pad position group to obtain the sensor pads;

[0116] Summarize the sensor pads to obtain multiple sensor pads, and denote these multiple sensor pads as the sensor circuit board.

[0117] The sensor circuit board was cured using a reflow oven to obtain a cured circuit board;

[0118] Obtain the gold wire bonding process parameter set, perform gold wire bonding operation based on the gold wire bonding process parameter set and the solidified circuit board, and obtain the integrated circuit board.

[0119] In essence, an integrated environment refers to an environment where multiple sensor assemblies and printed circuit boards are integrated. A solder paste printer is a machine used to apply solder paste to printed circuit boards; for example, a fully automatic solder paste printer. A pick-and-place machine is a machine used to pick up target sensor assemblies and align them with target pads for placement; for example, a high-speed automatic pick-and-place machine. A reflow oven is a machine used to cure the sensor pads; for example, an eight-zone fully automatic reflow oven.

[0120] It should be explained that the process of applying solder paste to a printed circuit board using a printing press refers to using a printing press to evenly apply solder paste onto the printed circuit board. Solder paste is the material used to connect the temperature sensor and the printed circuit board. For example, applying gold-tin alloy solder paste onto a printed circuit board using a printing press results in a gold-tin alloy solder paste circuit board (i.e., a solder paste circuit board).

[0121] It should be understood that the process of importing the solder paste circuit board into the pick-and-place machine is to utilize the machine to identify multiple pad positions on the solder paste circuit board and to perform placement operations on multiple sensor groups and the solder paste circuit board. The circuit board image is an image obtained by capturing the solder paste circuit board using the pick-and-place machine. Performing contour extraction on the circuit board image refers to extracting contours by comparing the grayscale values ​​of the pad positions with the grayscale values ​​of other locations on the solder paste circuit board (excluding the pad positions). The contours of the pad positions and the contours of the other locations extracted from the circuit board image constitute the multiple contours to be confirmed. The temperature pad contours and liquid level pad contours refer to the pre-constructed contours of the pad positions for the temperature sensor and the liquid level sensor that need to be soldered on the solder paste circuit board.

[0122] For example, the preset outline of the pad position corresponding to the temperature sensor (temperature pad outline) is a square, and the preset outline of the pad position corresponding to the liquid level sensor (liquid level pad outline) is a rectangle. The temperature pad outline and the liquid level pad outline are compared with multiple outlines to be confirmed. The outlines to be confirmed that are the same as the temperature pad outline and the liquid level pad outline are confirmed as the target temperature pad outline and the target liquid level pad outline. If the square outline to be confirmed is the same as the temperature pad outline, then the square outline to be confirmed is confirmed as the target temperature pad outline. If the rectangular outline to be confirmed is the same as the liquid level pad outline, then the rectangular outline to be confirmed is confirmed as the target liquid level pad outline.

[0123] In detail, if the positions of the temperature pads and liquid level pads are not precisely aligned with the corresponding temperature and liquid level sensors, the temperature and liquid level sensors will not be fixed to the solder paste circuit board, leading to slight shaking and subsequent failure of the temperature and liquid level detection probe during use. Therefore, this invention, with preset temperature and liquid level pad outlines, uses a pre-built image recognition algorithm to extract the outline to be confirmed. This outline is then compared with the preset temperature and liquid level pad outlines to confirm the positions of the temperature and liquid level sensors in the circuit board image. This eliminates the need for manual calibration by operators before each temperature and liquid level detection probe fabrication, reducing human error, improving the alignment accuracy between the temperature and liquid level pad positions and the corresponding temperature and liquid level sensors, reducing the possibility of temperature and liquid level detection probe failure, and achieving automated fabrication of temperature and liquid level detection probes. The image recognition algorithm is existing technology and will not be described in detail here.

[0124] Specifically, the solder paste circuit board coordinate system refers to the pre-constructed image coordinate system of the solder paste circuit board. Based on the contours of multiple target temperature pads and multiple target liquid level pads, the coordinate positions of the temperature pads and liquid level pads in the image coordinate system of the circuit board image are determined. Then, the coordinate positions in the image coordinate system of the circuit board image are converted into corresponding coordinate positions in the physical coordinate system constructed with the solder paste circuit board through a pre-constructed affine transformation. This confirms the multiple pad position groups at corresponding positions on the solder paste circuit board, with the temperature pad position corresponding to the target temperature pad contour and the liquid level pad position corresponding to the target liquid level pad contour. Affine transformation is existing technology and will not be elaborated upon here.

[0125] Importantly, the temperature pad position refers to the location where the temperature sensor is soldered on the solder paste circuit board, and the liquid level pad position refers to the location where the liquid level sensor is soldered on the solder paste circuit board.

[0126] It should be noted that the "using a pick-and-place machine to pick up the target sensor assembly" refers to using a vacuum nozzle on the pick-and-place machine to pick up the target sensor. The "aligning and placing the picked-up target sensor assembly and the target pad position group" refers to using the pick-and-place machine to move the target sensor assembly picked up by the vacuum nozzle to a position directly opposite the target pad position group and maintaining a certain distance, then pressing the target sensor assembly towards the target pad position group, and using solder paste to adhere the target sensor assembly to the solder paste circuit board to form sensor pads. The sensor circuit board refers to a collection of multiple sensor pads.

[0127] It is clear that the printed circuit board is a complete circuit board that integrates multiple sensor groups. During the confirmation of the preparation environment, the printed circuit board is divided into areas corresponding to each sensor group, and within each area corresponding to a sensor group, temperature pads and liquid level pads of different shapes are preset according to the temperature sensor and liquid level sensor, respectively. Therefore, the sensor circuit board is also a complete circuit board.

[0128] It is understood that the process of using a reflow oven to perform a curing operation on the sensor circuit board refers to the process of using a reflow oven to evaporate the volatile solution in the solder paste of the sensor circuit board, changing the solder paste from liquid to solid, and obtaining a cured circuit board. The purpose of using a reflow oven to perform a curing operation on the sensor circuit board is to make the combination between multiple sensor groups and multiple pad position groups more secure, and to avoid the failure of the prepared temperature and liquid level detection probe due to loosening and displacement during use.

[0129] Importantly, the aforementioned acquisition of the gold wire bonding process parameter set refers to acquiring the set of optimal process parameters for the gold wire bonding process. Gold wire bonding is a process used to perform electrical interconnection operations on multiple sensor groups and the printed circuit board to obtain an integrated circuit board. Process parameters refer to the parameters required during the gold wire bonding process, such as the gold content of the gold wire (metal wire), the diameter of the gold wire, the bonding temperature, and the bonding pressure. The values ​​of process parameters such as the gold content, diameter, bonding temperature, and bonding pressure constitute a set of gold wire bonding process parameters.

[0130] In detail, this invention effectively avoids problems such as gold wire embrittlement and breakage by obtaining a suitable set of gold wire bonding process parameters, thus preventing the temperature and liquid level detection probe from failing due to embrittlement and breakage during use. The gold wire bonding operation based on the set of gold wire bonding process parameters and the cured circuit board refers to the process of establishing electrical signal connections between multiple sensor groups and the printed circuit board in the sensor circuit board according to the set of gold wire bonding process parameters.

[0131] Specifically, obtaining the gold wire bonding process parameter set includes:

[0132] A knowledge graph of gold wire bonding is constructed based on pre-built historical data of gold wire bonding;

[0133] Extract multiple process parameters, their value ranges, and constraint rules from the knowledge graph;

[0134] Extract a process parameter value range from multiple process parameter value ranges sequentially to obtain the target value range, and then perform the following operations on the target value range:

[0135] Based on the preset number of process parameters, the target value range is divided into intervals to obtain multiple value intervals. One value interval is extracted from the multiple value intervals in turn, and a numerical extraction operation is performed on the extracted value interval to obtain the process parameter values. The process parameter values ​​are summarized to obtain multiple process parameter values ​​corresponding to the target value range. The multiple process parameter values ​​are recorded as the target process scattered value group.

[0136] Summarize the target process scattered value groups to obtain multiple target process scattered value groups corresponding to multiple process parameter value ranges. Perform parameter classification operation based on element-distinctness on multiple target process scattered value groups based on multiple process parameters to obtain multiple initial process parameter sets. The initial process parameter sets include multiple initial process parameters, and the initial process parameters correspond one-to-one with the process parameters.

[0137] If multiple initial process parameter sets all meet the constraint rules, then the multiple initial process parameter sets are confirmed as multiple process parameter sets.

[0138] The gold wire bonding process parameter set is obtained based on multiple process parameter sets.

[0139] It is understood that the aforementioned gold wire bonding historical data refers to data from the historical production process of gold wire bonding for testing probes, including but not limited to data on the process parameters of gold wire bonding when the produced temperature and level detection probes were usable, data on the process parameters of gold wire bonding when the produced temperature and level detection probes were not usable, and performance data of the temperature and level detection probes. The aforementioned knowledge graph refers to a knowledge graph of gold wire bonding constructed based on the historical gold wire bonding data. The knowledge graph includes the value ranges and constraint rules of multiple process parameters for gold wire bonding. Constructing a knowledge graph is existing technology, and will not be elaborated upon here.

[0140] It should be explained that extracting multiple process parameters, their value ranges, and constraint rules from the knowledge graph refers to identifying the process parameters in the gold wire bonding process from the knowledge graph, and obtaining the corresponding value ranges and constraint rules based on these identified parameters. For example, the process parameters identified from the knowledge graph include the gold content, diameter, bonding temperature, and bonding pressure of the gold wire. Then, the value ranges for the gold content and diameter of the gold wire are obtained from the knowledge graph. Constraint rules refer to various constraints on the values ​​of these process parameters, including but not limited to the following: with a fixed gold content, the bonding temperature cannot exceed a set temperature threshold. If the temperature exceeds the threshold, the gold wire will become brittle and break, causing the temperature and level detection probe to fail. Generally, the value ranges of these multiple process parameters are extracted from the data of the gold wire bonding process parameters when the produced temperature and level detection probe is functioning normally.

[0141] In detail, the target value range refers to the range of process parameter values ​​extracted from multiple process parameter value ranges. The number of process parameter sets refers to the number of multiple process parameter sets preset by the user. The number of process parameter sets is also the number of multiple candidate process parameter sets. The purpose of setting the number of process parameter sets is to limit the number of candidate process parameters generated when generating multiple candidate process parameter sets based on multiple process parameter sets, thereby avoiding the unlimited growth of the candidate process parameter sets. The interval segmentation of the target value range based on the preset number of process parameter sets refers to dividing the target value range into multiple value intervals on an average basis according to the number of process parameter sets, wherein the number of multiple value intervals is the number of process parameter sets. For example, if there is a target value range [0,3] and the preset number of process parameter sets is 3, then the target value range is divided into 3 value intervals on an average basis: [0,1], (1,2], and (2,3).

[0142] It should be understood that performing a numerical extraction operation on the extracted value range refers to randomly extracting a value from the extracted value range as a process parameter value. For example, if there are three value ranges [0,1], (1,2], and (2,3], and a process parameter value is randomly selected from each of the three value ranges, the three process parameter values ​​are 0.5, 1.3, and 2.6, respectively. Then, the multiple process parameter values ​​corresponding to the target value range are 0.5, 1.3, and 2.6.

[0143] It should be noted that the parameter classification operation based on element-distinctness refers to extracting a process parameter value corresponding to a process parameter from each of the multiple target process scatter sets. After the process parameter value corresponding to the current target process scatter set is extracted and an initial process parameter set is constructed, the process parameter value is not extracted again when constructing other initial process parameter sets. This ensures that each initial process parameter in the constructed initial process parameter set corresponds to a different process parameter, and the process parameter value corresponding to the same process parameter in different initial process parameter sets is different.

[0144] For example, given the following process parameters: gold content of the gold wire, and multiple process parameters corresponding to the gold content taking values ​​of 1, 2, and 3, the target process value set for the gold content is {1, 2, 3}. Given the following process parameters: diameter of the gold wire, and multiple process parameters for the diameter taking values ​​of 3, 4, and 5, the target process value set for the diameter is {3, 4, 5}. Given the following process parameters: bonding temperature of the gold wire, and multiple process parameters for the bonding temperature taking values ​​of 7, 8, and 9, the target process value set for the bonding temperature is {7, 8, 9}.

[0145] For a further example, if the gold content of the extracted gold wire is 1, and an initial process parameter set of {1, 3, 7} is constructed based on the gold content of the gold wire being 1, then when constructing other initial process parameter sets, the process parameter value of the gold content of the gold wire being 1 will not be extracted. By analogy, multiple initial process parameter sets of {1, 3, 7}, {2, 4, 8}, and {3, 5, 9} can be constructed, as can multiple initial process parameter sets of {1, 4, 9}, {2, 3, 8}, and {3, 5, 7}. Furthermore, the present invention uses constraint rules to check whether multiple initial process parameter sets are reasonable.

[0146] It is clear that when multiple initial process parameter sets contain values ​​that do not conform to the constraint rules, appropriate process parameter values ​​are reselected from the range of process parameter values ​​according to the constraint rules. These reselected values ​​are then used to replace the non-compliant initial process parameter sets. For example, if the gold content of the gold wire is 3, the gold wire bonding temperature cannot exceed the set temperature threshold of 6. However, the initial process parameter set contains a gold wire bonding temperature of 7. Therefore, a value no higher than 6 is randomly selected from the range of gold wire bonding temperature values ​​to replace the 7 value. Thus, the multiple process parameter sets are considered to be initial process parameter sets that conform to the constraint rules.

[0147] Furthermore, the step of obtaining the gold wire bonding process parameter set based on multiple process parameter sets includes:

[0148] Multiple candidate process parameter sets are generated based on multiple process parameter sets;

[0149] Simulation operations are performed on multiple candidate process parameter sets to obtain multiple simulated integrated circuit boards, wherein the simulated integrated circuit boards and candidate process parameter sets correspond one-to-one;

[0150] A board splitting operation is performed on multiple simulated integrated circuit boards to obtain multiple simulated testing component sets. Each simulated testing component set includes multiple simulated testing components, and each simulated testing component set corresponds one-to-one with a simulated integrated circuit board.

[0151] Performance prediction is performed on multiple simulation detection component sets to obtain multiple performance prediction sets. Each performance prediction set includes multiple performance prediction groups. There is a one-to-one correspondence between the simulation detection component sets and the performance prediction sets. Each performance prediction group includes multiple performance prediction values, and there is a one-to-one correspondence between the performance prediction groups and the simulation detection components.

[0152] Multiple performance qualities are calculated based on multiple performance prediction sets and pre-built performance quality calculation formulas. The multiple performance qualities are compared with the pre-built performance quality threshold. If there is a performance quality among the multiple performance qualities that is greater than or equal to the performance quality threshold, the performance quality with the largest value among the multiple performance qualities is extracted to obtain the maximum performance quality.

[0153] The candidate process parameter set corresponding to the maximum performance quality is determined from the multiple candidate process parameter sets to obtain the gold wire bonding process parameter set;

[0154] If multiple performance quality values ​​are all less than the performance quality threshold, then multiple candidate process parameter sets are used as the multiple process parameter sets, and the step of generating multiple candidate process parameter sets based on multiple process parameter sets is returned until the gold wire bonding process parameter set is obtained.

[0155] It should be noted that the generation of multiple candidate process parameter sets based on multiple process parameter sets refers to generating multiple candidate process parameter sets using a genetic algorithm based on multiple process parameter sets. These multiple process parameter sets are obtained from a knowledge graph constructed based on the historical data of gold thread bonding; that is, they are process parameter sets obtained based on historical experience and are not necessarily the optimal set. Therefore, this invention uses a genetic algorithm to generate multiple unknown candidate process parameter sets. Generally, these multiple candidate process parameter sets are all process parameter sets that conform to the constraint rules, used to explore a better set of process parameters. Genetic algorithms are existing technology and will not be described in detail here.

[0156] It is understood that performing simulation operations on multiple candidate process parameter sets refers to the process of simulating wire bonding to obtain multiple simulated integrated circuit boards using simulation software (e.g., Ansys HFSS software) based on multiple candidate process parameter sets. Each candidate process parameter generates a simulated integrated circuit board.

[0157] Specifically, the process of performing a board splitting operation on multiple simulated integrated circuit boards refers to sequentially extracting one simulated integrated circuit board from the multiple simulated integrated circuit boards and dividing the simulated integrated circuit board based on the multiple sensor groups to obtain multiple simulated detection component sets. For example, if there are 3 simulated integrated circuit boards, and each simulated integrated circuit board has 4 simulated sensor groups, then the 3 simulated integrated circuit boards are divided into 3 simulated detection component sets, and each of the 3 simulated detection component sets contains 4 simulated detection components.

[0158] It is clear that performing performance prediction on multiple sets of simulated detection components refers to the process of predicting the performance of multiple sets of simulated detection components based on a pre-built performance prediction model, thereby obtaining multiple performance prediction sets. Generally, the performance prediction model is trained and tested using multiple sets of detection components and multiple performance sets from the historical data of gold wire bonding. The performance prediction model can predict the corresponding multiple performance prediction sets based on the numerical values ​​of the physical properties of the multiple input simulated detection components. The performance prediction model is a neural network.

[0159] It needs to be explained that multiple detection component sets refer to the set of physical structure values ​​of multiple detection components in the actual production process, while multiple detection performance sets refer to the set of performance values ​​obtained from actual performance tests of multiple detection components in the actual production process. The above-mentioned training based on multiple detection component sets and multiple detection performance sets from the historical data of gold wire bonding refers to: using multiple detection component sets and multiple performance prediction sets predicted by the performance prediction model as input features, using multiple detection performance sets as true labels, and minimizing the mean absolute error between the multiple performance prediction sets and the true labels using the Adam algorithm, thereby completing the training of the neural network. This training process is a common operation in neural network training and will not be elaborated further.

[0160] In detail, the multiple predicted performance values ​​refer to the predicted values ​​of the target simulation detection component in multiple performance types obtained by performing performance prediction on the target simulation detection component. The performance prediction set includes multiple performance prediction groups, and each performance prediction group includes multiple performance prediction values ​​of a simulation detection component in a simulation detection component set. For example, by performing performance prediction on performance types such as signal strength, connection strength (stability of the connection between the liquid level sensor and the solidified circuit board), and sensitivity (accuracy of temperature and liquid level detection) of the simulation detection component, multiple predicted performance values ​​are obtained: signal strength 6, connection strength 7, and sensitivity 8. Then, the performance prediction group corresponding to a simulation detection component (i.e., a sensor) is: {signal strength 6, connection strength 7, sensitivity 8}.

[0161] It should be noted that the present invention quantifies the performance of the detection component through a performance quality calculation formula, and compares the calculated performance quality with a performance quality threshold to determine whether the detection component meets the requirements of multiple performance types. Therefore, the performance quality threshold refers to the performance quality threshold of the temperature and liquid level detection probe set by the user, which can be set by the average performance quality of multiple normally used temperature and liquid level detection probes in actual production.

[0162] It should be understood that when the performance quality is greater than the performance quality threshold, it indicates that the simulation test component set corresponding to the performance quality meets the requirements of multiple performance types in the simulation, which means that the candidate process parameter set represented by the simulation test component set can produce qualified test components in the simulation. At the same time, the higher the performance quality, the better the performance of the simulation test component set produced by the corresponding candidate process parameter set compared with other candidate process parameter sets. Therefore, in this invention, when one or more performance qualities are greater than or equal to the performance quality threshold, the candidate process parameter set corresponding to the maximum value of one or more performance qualities is extracted as the gold wire bonding process parameter set.

[0163] It is understood that the multiple performance qualities obtained from the multiple candidate process parameter sets are all simulation results. Therefore, after performing the gold wire bonding operation based on the gold wire bonding process parameter set and the cured circuit board to obtain the integrated circuit board, the integrated circuit board needs to be subjected to actual performance testing. Only when the performance quality of the prepared integrated circuit board is greater than or equal to the performance quality threshold can the prepared integrated circuit board be confirmed as a qualified integrated circuit board, and the subsequent step of applying adhesive to the integrated circuit board can be carried out.

[0164] Furthermore, the calculation of multiple performance qualities based on multiple performance prediction sets and pre-built performance quality calculation formulas includes:

[0165] Extract a performance prediction set from multiple performance prediction sets sequentially to obtain the target performance prediction set, and then perform the following operations on the target performance prediction set:

[0166] For each performance prediction group in the target performance prediction group set, a vector construction operation is performed to obtain multiple prediction performance vectors, and a prediction performance matrix is ​​generated based on the multiple prediction performance vectors.

[0167] Multiple performance types are identified based on the predicted performance matrix, multiple weights are assigned to the multiple performance types, and a weight vector is constructed using the multiple weights.

[0168] Calculate the weight performance matrix based on the weight vector and the prediction performance matrix;

[0169] Normalize the weight performance matrix to obtain the normalized performance matrix, as shown below:

[0170] ,

[0171] in, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the performance prediction group An index of normalized performance values, Indicates the first in the target performance prediction group set Indexes for performance prediction groups;

[0172] By summing the normalized performance matrices, multiple normalized performance matrices are obtained;

[0173] Multiple performance qualities are calculated based on multiple normalized performance matrices and performance quality calculation formulas.

[0174] It is understood that the target performance prediction set refers to the performance prediction set extracted from multiple performance prediction sets. The vector construction operation performed on each performance prediction set in the target performance prediction set refers to the process of constructing a row vector from the multiple predicted performance values ​​in each performance prediction set. For example, if there are multiple predicted performance values: signal strength 6, connection strength 7, and sensitivity 8, then the predicted performance vector is... The prediction performance matrix is ​​the matrix obtained by constructing column vectors from each of the multiple prediction performance vectors as word blocks.

[0175] For example, given a first simulation detection component, a second simulation detection component set, and a third simulation detection component, the corresponding performance prediction groups are: a first performance prediction group for the first simulation detection component {signal strength 6, connection strength 7, sensitivity 8}, a second performance prediction group for the second simulation detection component {signal strength 5, connection strength 2, sensitivity 8}, and a third performance prediction group for the third simulation detection component {signal strength 7, connection strength 7, sensitivity 7}. The corresponding prediction performance matrix is ​​as follows: ,in, This represents the predicted performance vector of the first simulation testing component. Generally, each predicted performance matrix corresponds to a simulated integrated circuit board based on a candidate process parameter.

[0176] It should be understood that identifying multiple performance types based on the prediction performance matrix refers to determining the multiple performance types corresponding to multiple predicted performance values ​​in multiple prediction performance matrices. For example, if the prediction performance matrix includes multiple predicted performance values ​​6, 7, and 8, where 6, 7, and 8 correspond to signal strength, connection strength, and sensitivity, respectively, then the multiple performance types are signal strength, connection strength, and sensitivity. Assigning multiple weights to multiple performance types refers to assigning a weight to each performance type using an expert scoring method. For example, if the multiple performance types are signal strength, connection strength, and sensitivity, and the expert scoring method assigns weights of 0.5, 0.2, and 0.3 to signal strength, connection strength, and sensitivity, then the weight vector is... The expert scoring method is existing technology, and will not be described further in this invention.

[0177] Importantly, the calculation of the weight performance matrix based on the weight vector and the prediction performance matrix refers to performing a matrix expansion of the weight vector with the same shape as the first prediction performance matrix to obtain the weight matrix, and then performing a Hadamard product operation on the weight matrix and the prediction performance matrix to obtain the weight performance matrix. The matrix expansion with the same shape as the first prediction performance matrix means using the weight vector as a sub-block and copying it as a sub-block weight vector to construct a matrix with the same shape as the first prediction performance matrix.

[0178] For example, there exists a first prediction performance matrix. and weight vector The weight vector is expanded using a matrix of the same shape based on the first prediction performance matrix to obtain the weight matrix. Perform the Hadamard product operation on the weight matrix and the first prediction performance matrix (i.e.: ,in, (using the Hadamard product notation) to obtain the first weighted performance matrix. .

[0179] It is understandable that the multiple performance prediction values ​​in each column of the prediction performance matrix correspond to the same performance type as the weights of the corresponding columns in the weight vector. Therefore, by performing a matrix expansion of the weight vector based on the same type as the first prediction performance matrix, a weight matrix is ​​obtained. Then, a Hadamard product operation is performed on the weight matrix and the prediction performance matrix to make the resulting weight performance matrix a weighted prediction performance matrix.

[0180] Specifically, the normalization operation on the weight performance matrix refers to the process of normalizing multiple predicted performance values ​​multiplied by the weight vector in the weight performance matrix using vector normalization to obtain a normalized performance matrix. Vector normalization is existing technology and will not be elaborated upon here.

[0181] In detail, the calculation of multiple performance qualities based on multiple normalized performance matrices and performance quality calculation formulas includes:

[0182] Extract one of the multiple performance types sequentially to obtain the target performance type, and perform the following operations on the target performance type:

[0183] Extract multiple normalized performance values ​​corresponding to the target performance type from the multiple normalized performance matrices, and identify the normalized performance value with the largest value and the normalized performance value with the smallest value among the multiple normalized performance values ​​to obtain the maximum performance value and the minimum performance value.

[0184] The maximum and minimum performance values ​​are summarized separately to obtain the maximum performance value vector and minimum performance value vector corresponding to multiple performance types. The maximum performance value vector includes multiple maximum value elements, and each maximum value element corresponds one-to-one with the performance type.

[0185] Extract one of the multiple normalized performance matrices sequentially to obtain the target normalized performance matrix. Perform the following operations on the target normalized performance matrix:

[0186] Based on the multiple simulation detection component sets, multiple normalized performance vectors are obtained from the target normalized performance matrix. Then, one normalized performance vector is extracted from the multiple normalized performance vectors to obtain the target performance vector. The following operations are performed on the target performance vector:

[0187] Calculate the Euclidean distance between the target performance vector and the maximum performance value vector to obtain the positive ideal distance; calculate the Euclidean distance between the target performance vector and the minimum performance value vector to obtain the negative ideal distance.

[0188] Performance similarity is calculated based on positive and negative ideal distances;

[0189] Summarize the performance similarity scores to obtain multiple performance similarity scores, and construct the performance similarity vector corresponding to the target normalized performance matrix based on the performance similarity scores;

[0190] By summing the performance proximity vectors, multiple performance proximity vectors corresponding to multiple performance prediction sets are obtained, where there is a one-to-one correspondence between the performance prediction sets and the performance proximity vectors.

[0191] Multiple performance qualities are calculated based on multiple performance proximity vectors and performance quality calculation formulas.

[0192] It should be noted that the target performance type refers to the performance type extracted from multiple performance types. The maximum performance value vector is the set of the maximum performance values ​​of multiple performance types, and the minimum performance value vector is the set of the minimum performance values ​​of multiple performance types.

[0193] For example, a first normalized performance matrix exists as follows: ,

[0194] There exists a second normalized performance matrix as shown below: ,

[0195] In this matrix, the first column of both the first and second normalized performance matrices corresponds to signal strength, the second column corresponds to connectivity strength, and the third column corresponds to sensitivity. Therefore, the normalized performance values ​​extracted based on signal strength are {0.1, 0.1, 0.2, 0.1, 0.2, 0.2}, the normalized performance values ​​extracted based on connectivity strength are {0.3, 0.4, 0.3, 0.5, 0.4, 0.3}, and the normalized performance values ​​extracted based on sensitivity are {0.5, 0.6, 0.7, 0.5, 0.5, 0.4}. The maximum and minimum performance values ​​corresponding to signal strength are 0.2 and 0.1, respectively; the maximum and minimum performance values ​​corresponding to connectivity strength are 0.5 and 0.3, respectively; and the maximum and minimum performance values ​​corresponding to sensitivity are 0.7 and 0.4, respectively. The maximum performance value vector is... The minimum performance value vector is .

[0196] Specifically, obtaining multiple normalized performance vectors from the target normalized performance matrix based on the multiple simulation detection component sets refers to determining the simulation detection component set corresponding to the target normalized performance matrix in the multiple simulation detection component sets, and extracting multiple normalized performance vectors based on the multiple simulation detection components in the corresponding simulation detection component set. Each of the multiple normalized performance vectors is a row vector of the target normalized performance matrix.

[0197] For example, continuing with the previous method, a first normalized performance matrix and a second normalized performance matrix exist. The first normalized performance matrix is ​​calculated based on the first simulation detection component set, and the second normalized performance matrix is ​​calculated based on the second simulation detection component set. The first normalized performance matrix is ​​extracted as the target normalized performance matrix. Since the first simulation detection component set contains a first simulation detection component, a second simulation detection component, and a third simulation detection component, the first row of the first normalized performance matrix corresponds to the normalized performance values ​​of the first simulation detection component in the first simulation detection component set for three performance types: signal strength, connection strength, and sensitivity. The second row of the first normalized performance matrix corresponds to the normalized performance values ​​of the second simulation detection component in the first simulation detection component set for three performance types: signal strength, connection strength, and sensitivity. The third row of the first normalized performance matrix corresponds to the normalized performance values ​​of the third simulation detection component in the first simulation detection component set for three performance types: signal strength, connection strength, and sensitivity. Furthermore, the plurality of normalized performance vectors are respectively the first normalized performance vectors of the first simulation detection component. The second normalized performance vector of the second simulation detection component The third normalized performance vector of the third simulation detection component Similarly, the second normalized performance matrix can be used as the target normalized performance matrix to obtain multiple normalized performance vectors, which will not be elaborated further in this invention.

[0198] It is clear that the target performance vector is a normalized performance vector extracted from multiple normalized performance vectors. The positive ideal distance refers to the Euclidean distance between the target performance vector and the vector of maximum performance value, while the negative ideal distance refers to the Euclidean distance between the target performance vector and the vector of minimum performance value. Euclidean distance is existing technology and will not be elaborated upon here.

[0199] Importantly, the calculation of performance proximity based on positive and negative ideal distances refers to substituting the positive and negative ideal distances into a pre-constructed performance proximity calculation formula to calculate the performance proximity. The performance proximity calculation formula is as follows:

[0200] ,

[0201] in, Indicates performance similarity. Indicates the negative ideal distance. This represents the positive ideal distance. A greater performance proximity indicates that the target performance vector is closer to the maximum performance value vector. This signifies that the performance of the simulation testing component corresponding to the target performance vector is closer to the optimal performance, and consequently, the candidate process parameter set corresponding to the corresponding simulation testing component is closer to the optimal process parameter set. Generally, according to the performance proximity formula, the performance proximity range is between 0 and 1.

[0202] Furthermore, the calculation of multiple performance qualities based on multiple performance proximity vectors and performance quality calculation formulas includes:

[0203] Extract one of the multiple normalized performance matrices sequentially to obtain the key normalized performance matrix. Perform the following operations on the key normalized performance matrix:

[0204] Based on the multiple performance types, multiple key performance vectors are obtained from the key normalized performance matrix. Each of the multiple key performance vectors is the transpose of the column vector of the key normalized performance matrix, and the key performance vectors correspond one-to-one with the performance types. Each key performance vector includes multiple key performance values.

[0205] Extract one key performance vector from multiple key performance vectors sequentially to obtain the target key performance vector, and then perform the following operations on the target key performance vector:

[0206] Calculate the standard deviation of multiple key performance values ​​to obtain the performance standard deviation;

[0207] Summarize the performance standard deviations to obtain multiple performance standard deviations corresponding to multiple performance types, and construct a performance standard deviation vector based on the multiple performance standard deviations;

[0208] The performance proximity vector corresponding to the normalized performance matrix represented by the performance standard deviation vector is extracted from multiple performance proximity vectors to obtain a definite proximity vector. The performance quality is then calculated based on the performance standard deviation vector, the definite proximity vector, and the performance quality calculation formula.

[0209] By summarizing the performance qualities, multiple performance qualities are obtained, each of which corresponds one-to-one with a set of candidate process parameters.

[0210] It should be noted that the specific operation of obtaining multiple key performance vectors from the key normalized performance matrix based on the multiple performance types is as follows: extract the transpose of the column vector corresponding to each performance type from the multiple performance types in the key normalized performance matrix, and the transpose of the resulting multiple column vectors is the key performance vector.

[0211] For example, there exists a first normalized performance matrix. Second normalized performance matrix The performance types include signal strength, connection strength, and sensitivity. The first normalized performance matrix is ​​extracted as the key normalized performance matrix. The key performance vector corresponding to signal strength is then the transpose of the column vector represented by the first column of the first normalized performance matrix. Similarly, the key performance vector corresponding to connection strength is: The key performance vector corresponding to sensitivity is Further extract the key performance vectors corresponding to the signal strength. As the target key performance vector.

[0212] It is understood that the performance standard deviation refers to the standard deviation of multiple key performance values. The performance standard deviation vector refers to the vector constructed by multiple performance standard deviations corresponding to multiple key performance vectors. Standard deviation is prior art, and will not be elaborated upon here.

[0213] It should be explained that the determined proximity vector refers to the performance proximity vector corresponding to the normalized performance matrix represented by the performance standard deviation vector among the multiple performance proximity vectors. For example, given a first normalized performance matrix and a second normalized performance matrix, the first performance proximity vector and the second performance proximity vector can be calculated from the first normalized performance matrix and the second normalized performance matrix, respectively. If the key normalized performance matrix is ​​the first normalized performance matrix, and the first performance standard deviation vector can be calculated from the key normalized performance matrix, then the first performance proximity vector among the first performance proximity vector and the second performance proximity vector is the determined proximity vector.

[0214] It is understood that the calculation of performance quality based on the performance standard deviation vector, the determination of proximity vector, and the performance quality calculation formula refers to substituting the performance standard deviation vector and the determination of proximity vector into the performance quality calculation formula to calculate the performance quality. The calculated performance quality is derived from the normalized performance matrix corresponding to the performance standard deviation vector and the determination of proximity vector. Therefore, it can characterize the performance quality of the simulation test component set corresponding to the normalized performance matrix, and further characterize the merits of the candidate process parameter set corresponding to the simulation test component set (the greater the performance proximity, the closer the corresponding simulated integrated circuit board is to the optimal performance, and the closer the candidate process parameter set is to the optimal process parameter set).

[0215] Furthermore, the performance quality calculation formula is as follows:

[0216] ,

[0217] in, Indicates performance quality. Indicates the weight of performance similarity. Indicates the standard deviation weight of performance. Indices representing multiple performance similarity values ​​in the performance similarity vector. This represents the first of several proximity values ​​in the performance proximity vector. A proximity index, The first element in the performance proximity vector represents the... Performance similarity Indices representing multiple performance standard deviations of the performance standard deviation vector. The first of several performance standard deviations in the performance standard deviation vector represents the first of the several performance standard deviations. An index with a performance standard deviation, The first element in the performance standard deviation vector represents the... One performance standard deviation.

[0218] It should be noted that the performance proximity weight refers to the weight of the performance proximity vector preset by the user, and the performance standard deviation weight refers to the weight of the performance standard deviation vector preset by the user. Optionally, the performance proximity weight and the performance standard deviation weight are respectively... The performance proximity vector includes the performance proximity of each simulation testing component in the simulation testing component set. The performance standard deviation vector includes the performance standard deviation of each performance type in the simulation testing component set across multiple performance types. The performance standard deviation represents the consistency of multiple simulation testing components in the simulation testing component set across the same performance type. Therefore, this invention can characterize the consistency of the simulation testing component set across multiple performance types through the performance standard deviation vector. The smaller the performance standard deviation, the higher the consistency of multiple simulation testing components in the simulation testing component set across the same performance type.

[0219] It should be explained that consistency refers to the difference in performance prediction values ​​of multiple simulation testing components generated from a candidate process parameter set for the same performance type. The greater the difference, the lower the consistency, and the more unstable the candidate process parameter set is when simulating the production of multiple simulation testing components, resulting in large fluctuations in the performance prediction values ​​of multiple simulation testing components for the same performance type.

[0220] For example, given multiple performance types—signal strength, connection strength, and sensitivity—performance prediction is performed on a first, second, and third simulation detection component manufactured using the same candidate process parameter set to obtain a normalized performance matrix. Among them, the signal strength corresponds to the key performance vector of the normalized performance matrix. Similarly, the key performance vector corresponding to connection strength is: The key performance vector corresponding to sensitivity is Among them, the key performance vector corresponding to signal strength. The first column of the code corresponds to the first simulation detection component, the second column corresponds to the second simulation detection component, and the third column corresponds to the third simulation detection component.

[0221] As a further example, the key performance vector corresponding to the normalized performance matrix of the signal strength is calculated. The standard deviation of the signal strength is used to obtain the performance standard deviation corresponding to the signal strength. Since the first, second, and third simulation detection components are all produced using the same candidate process parameter set, the performance standard deviation corresponding to the signal strength should theoretically approach 0. A larger performance standard deviation indicates greater differences in performance values ​​among the multiple simulation detection components produced by the corresponding candidate process parameter set, suggesting large performance fluctuations. This indicates that the corresponding candidate process parameter set cannot produce multiple simulation detection components with stable performance, making it unreliable and unsuitable for selection as the process parameter set for gold wire bonding.

[0222] Furthermore, this invention aims to assess the consistency and performance quality of simulated integrated circuit boards fabricated from candidate process parameter sets across multiple performance types by using performance standard deviation vectors and performance proximity vectors as parameters for performance quality calculation. Simultaneously, this invention uses preset performance proximity weights and preset performance standard deviation weights to find process parameter sets that meet the expected performance type requirements in terms of consistency and performance quality across multiple performance types. For example, if the requirement for performance type consistency is high, it can be achieved by increasing the performance standard deviation weight and decreasing the performance proximity weight. Conversely, if the requirement for performance quality across multiple performance types is high, it can be achieved by decreasing the performance standard deviation weight and increasing the performance proximity weight.

[0223] Furthermore, the step of performing gold wire bonding operation based on the gold wire bonding process parameter set and the cured circuit board to obtain an integrated circuit board includes:

[0224] The gold wire bonding process parameter set is imported into the pre-built gold wire bonding machine to obtain the setting bonding machine. The setting bonding machine is used to confirm the positions of multiple sensor groups and multiple bonding position groups in the cured circuit board. Among them, the bonding position group includes a temperature bonding position and a liquid level bonding position.

[0225] The location of a sensor group is determined sequentially from multiple sensor group locations to obtain the target sensor group location. The following operations are then performed on the target sensor group location:

[0226] Temperature metal wire and liquid level metal wire are obtained, and a discharge operation is performed on the temperature metal wire and liquid level metal wire using a setting bonding machine to obtain temperature metal ball and liquid level metal ball.

[0227] The temperature metal ball and the liquid level metal ball are pressed together to the target sensor group position to obtain the temperature welding metal wire and the liquid level welding metal wire;

[0228] Determine the target binding position group corresponding to the target sensor group position among multiple binding position groups to obtain the target binding position group;

[0229] The temperature welding metal wire and the liquid level welding metal wire are welded to the target bonding position group to obtain the gold wire bonding pad.

[0230] The gold wire bonding pads are aggregated to obtain multiple gold wire bonding pads, which are then referred to as the integrated circuit board.

[0231] It is understood that importing the gold wire bonding process parameter set into the pre-built gold wire bonding machine refers to setting the gold wire bonding process parameter set in the gold wire bonding machine. The gold wire bonding machine refers to a machine used for gold wire bonding, such as a fully automatic gold wire bonding machine. The process of using the bonding machine to confirm the positions of multiple sensor groups and multiple bonding position groups in the cured circuit board is consistent with the process of using a pick-and-place machine to perform image acquisition operations on the solder paste circuit board and determining the corresponding multiple pad position groups in the solder paste circuit board; these will not be elaborated upon here. Temperature bonding position refers to the position where the temperature sensor performs gold wire bonding, and liquid level bonding position refers to the position where the liquid level sensor performs gold wire bonding. Sensor group position refers to the position of the sensor group obtained after performing alignment and placement operations on the target sensor group and the target pad position group. Target sensor group position refers to the position of one sensor group extracted from multiple sensor group positions.

[0232] It should be explained that the temperature metal wire is a metal wire used for gold wire bonding of the temperature sensor, for example, a gold wire with a gold content of 99.98% and a diameter of 25 micrometers. The liquid level metal wire is a metal wire used for gold wire bonding of the liquid level sensor. The temperature metal ball is obtained by discharging one end of the temperature metal wire using a bonding machine, causing that end of the temperature metal wire to heat up and melt into a spherical shape. The process for obtaining the liquid level metal ball is the same as that for obtaining the temperature metal ball, and will not be described again here. Generally speaking, the solidified circuit board is a complete circuit board. The solidified circuit board is divided into areas corresponding to the gold wire bonding of each sensor group, and in the gold wire bonding area corresponding to each sensor group, different temperature bonding positions and liquid level bonding positions are marked according to the temperature sensor and liquid level sensor, respectively.

[0233] It should be understood that the process of pressing the temperature metal ball and the liquid level metal ball to the target sensor group position includes: aligning the temperature metal ball and the liquid level metal ball to the positions of the temperature sensor and the liquid level sensor corresponding to the target sensor group position and maintaining a distance (e.g., 1 cm), and then applying pressure to the temperature metal ball and the liquid level metal ball so that the temperature metal ball and the liquid level metal ball are connected to the temperature sensor and the liquid level sensor respectively, to obtain the temperature welding metal wire and the liquid level welding metal wire.

[0234] Specifically, to ensure that the temperature and liquid level metal wires between multiple sensor groups do not cross during the gold wire bonding process, thus facilitating subsequent board separation operations, each sensor group position corresponds to a bonding position group. The target bonding position group refers to the bonding position group corresponding to the target sensor group position among the multiple bonding position groups.

[0235] Importantly, the process of welding the temperature welding metal wire and the liquid level welding metal wire to the target bonding position group refers to performing a discharge operation on the other end of the temperature welding metal wire (the end that has not been formed into a ball and pressed to the temperature sensor) and the other end of the liquid level welding metal wire, and pressing them to the corresponding temperature bonding position and liquid level bonding position in the target bonding position group, thereby obtaining the gold wire bonding pad.

[0236] S3. Apply adhesive to the integrated circuit board to obtain an adhesive circuit board. Separate the adhesive circuit board based on multiple sensor groups to obtain multiple detection components, wherein the sensor groups and detection components correspond one-to-one.

[0237] Understandably, an adhesive circuit board refers to a circuit board after adhesive has been applied to an integrated circuit board. It is used to fix multiple detection components to multiple probe housings. For example, applying epoxy resin adhesive to an integrated circuit board yields an epoxy resin circuit board (i.e., an adhesive circuit board). The process of splitting the adhesive circuit board based on multiple sensor groups refers to cutting the adhesive circuit board according to the multiple sensor groups to obtain multiple detection components. For example, if there are four sensor groups, and each sensor group can be used to prepare a temperature and liquid level detection probe, then the adhesive circuit board is cut into four detection components using a cutting tool based on the four sensor groups. Each detection component contains one temperature sensor and one liquid level sensor.

[0238] S4. Sequentially extract one detection component from multiple detection components to obtain the target detection component. Perform the following operations on the target detection component: extract the probe housing corresponding to the detection component from multiple probe housings to obtain the target probe housing. Align the target detection component with the target probe housing and perform a pressure operation to obtain the component housing. Perform a heat curing operation on the component housing to obtain the cured detection component.

[0239] It should be noted that the target detection component refers to one of the multiple detection components extracted from the target detection component. The target probe housing refers to the probe housing corresponding to the target detection component among the multiple probe housings, which is determined based on the pre-built probe housing number and the detection component number. For example, detection component number 1 corresponds to probe housing number 1.

[0240] It should be understood that aligning the target detection component with the target probe housing and applying pressure refers to moving the target detection component to a position directly opposite the target probe housing while maintaining a certain distance, and then pressing the target detection component against the target probe housing, adhering the target detection component to the target probe housing using the adhesive to obtain the component housing. The heating and curing operation of the component housing refers to heating the component housing using a heater (e.g., a hot air circulating oven), causing the volatile solvents in the adhesive of the component housing to evaporate, thereby completing the curing process to obtain a cured detection component.

[0241] Importantly, the purpose of performing the heat curing operation on the component housing is to convert the adhesive into a solid adhesive, thereby fixing the detection component to the probe housing and preventing the temperature and level detection probe from becoming loose or displaced during use.

[0242] S5. Identify the signal cable corresponding to the curing detection component from multiple signal cables to obtain the target signal cable. Perform a welding operation on the curing detection component and the target signal cable to obtain the signal detection component. Perform a potting operation on the signal detection component to obtain the potted detection probe. Perform a heat curing operation on the potted detection probe to obtain the initial detection probe. Summarize the initial detection probes to obtain multiple initial detection probes.

[0243] In detail, the process of identifying the signal cable corresponding to the cured detection component from multiple signal cables is the same as the process of extracting the probe housing corresponding to the detection component from multiple probe housings, and will not be repeated here. Performing a soldering operation on the cured detection component and the target signal cable refers to the process of soldering (e.g., wires) the target signal cable to the cured detection component (e.g., soldering, laser soldering) to obtain the signal detection component. For the cured detection component to operate normally, current needs to be transmitted to it. Therefore, this invention allows the cured detection component to accept current transmitted from the external environment by soldering the signal cable to it.

[0244] Specifically, the potting operation on the signal detection component refers to the process of introducing potting compound into the probe housing, so that the leads and integrated circuit board are completely encapsulated by the potting compound to form a potted detection probe. The purpose of the potting operation on the signal detection component is to fix the leads and integrated circuit board. The heat curing operation on the potted detection probe is the same as the heat curing operation on the component housing in terms of process and beneficial effects, and will not be described again here.

[0245] S6. Perform quality control screening operations based on detection performance testing and packaging inspection on multiple initial detection probes to obtain multiple temperature and liquid level detection probes.

[0246] It is understood that the quality control screening operation based on detection performance testing and packaging inspection of multiple initial detection probes refers to performing performance tests on multiple initial detection probes for temperature and liquid level detection, and performing packaging inspections on multiple initial detection probes. Then, based on the results of the performance tests and packaging inspections, multiple temperature and liquid level detection probes are selected. For example, the initial detection probes are used to measure temperature and liquid level, and the detection performance, such as measurement sensitivity, measurement accuracy, and signal stability, is monitored in real time during the measurement process to complete the detection performance test. The temperature and liquid level detection probes are the initial detection probes that have passed the detection performance test.

[0247] It should be understood that passing the performance test means that the initial test probe is deemed qualified when the sensitivity, measurement error, and stability standards all meet the corresponding performance standards. (For example, if the sensitivity standard is greater than or equal to the sensitivity standard, the measurement sensitivity is considered to meet the performance test standard. The same logic applies to other performance aspects, and further examples will not be provided here.) If any performance aspect fails to meet the performance test standard, it indicates that the measured temperature and liquid level are inaccurate, and the corresponding initial test probe is unqualified and should be discarded. Generally, only one or more initial test probes that pass the performance test can proceed to the packaging inspection.

[0248] Specifically, the encapsulation inspection refers to checking whether the appearance and packaging of one or more qualified initial test probes are damaged. If a qualified initial test probe is damaged, it will also be recorded as an unqualified initial test probe. Only the initial test probes that pass the encapsulation inspection and the test performance test will be used as temperature and liquid level test probes.

[0249] To address the problems described in the background art, multiple sensor groups and printed circuit boards are integrated to obtain an integrated circuit board. It is evident that this invention integrates multiple sensor groups onto a single printed circuit board, solving the problem of low integration density within the probe's internal structure. Simultaneously, this invention replaces the traditional error-prone manual soldering by accurately predicting the gold wire bonding process parameters, fundamentally ensuring high consistency in the electrical performance of each temperature and level detection probe. Furthermore, when the performance of the temperature and level detection probe fails to meet expectations, the gold wire bonding process parameters are corrected, resulting in an integrated temperature and level detection probe with superior performance and better stability. This invention integrates multiple traditionally serial and separate bonding and curing steps into a highly automated surface mount and gold wire bonding process, significantly improving production efficiency. Therefore, this invention predicts the gold wire bonding process parameters through simulation and corrects the parameters when expected performance is not achieved, thereby obtaining an integrated temperature and level detection probe with superior performance and better stability.

[0250] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0251] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An automated manufacturing method for an integrated temperature and liquid level detection probe, characterized in that, The method includes: The system receives a probe preparation instruction and confirms the preparation environment based on the instruction. The preparation environment includes multiple sensor groups, printed circuit boards, multiple probe housings, and multiple signal cables. The sensor groups include a temperature sensor and a liquid level sensor. An integration operation is performed on multiple sensor groups and printed circuit boards to obtain an integrated circuit board; An adhesive application operation is performed on the integrated circuit board to obtain an adhesive circuit board. Based on multiple sensor groups, a board separation operation is performed on the adhesive circuit board to obtain multiple detection components, wherein the sensor groups and detection components correspond one-to-one. Extract one detection component from multiple detection components sequentially to obtain the target detection component, and perform the following operations on the target detection component: Extract the probe housing corresponding to the detection component from multiple probe housings to obtain the target probe housing; Align the target detection component with the target probe housing and apply pressure to obtain the component housing. Then, perform a heat curing operation on the component housing to obtain the cured detection component. The signal cable corresponding to the curing detection component is identified from multiple signal cables to obtain the target signal cable. A soldering operation is then performed on the curing detection component and the target signal cable to obtain the signal detection component. A potting compound is applied to the signal detection component to obtain a potted detection probe. A heat curing operation is then performed on the potted detection probe to obtain an initial detection probe. The initial detection probes are summarized to obtain multiple initial detection probes; Multiple initial detection probes were subjected to quality control screening based on detection performance testing and packaging inspection to obtain multiple temperature and liquid level detection probes.

2. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 1, characterized in that, The process of integrating multiple sensor groups and printed circuit boards to obtain an integrated circuit board includes: Identify the integrated environment, which includes the printer, pick-and-place machine, and reflow oven; Solder paste is applied to the printed circuit board using a printing press to obtain a solder paste circuit board. The solder paste circuit board is then fed into a pick-and-place machine, which performs an image acquisition operation on the solder paste circuit board to obtain a circuit board image. A contour extraction operation is performed on the circuit board image to obtain multiple contours to be confirmed. A comparison operation is performed on the multiple contours to be confirmed using the pre-constructed temperature pad contour and the pre-constructed liquid level pad contour to obtain multiple target temperature pad contours and multiple target liquid level pad contours. Based on the pre-constructed solder paste circuit board coordinate system, multiple target temperature pad contours and multiple target liquid level pad contours, multiple pad position groups are determined in the solder paste circuit board. Each pad position group includes a temperature pad position and a liquid level pad position. The temperature pad position corresponds one-to-one with the target temperature pad contour, and the liquid level pad position corresponds one-to-one with the target liquid level pad. Extract one sensor group from multiple sensor groups sequentially to obtain the target sensor group, and perform the following operations on the target sensor group: The target pad position group is obtained by determining the pad position group corresponding to the target sensor group from multiple pad position groups; A pick-and-place machine is used to pick up the target sensor group, and an alignment and placement operation is performed on the picked-up target sensor group and the target pad position group to obtain the sensor pads; Summarize the sensor pads to obtain multiple sensor pads, and denote these multiple sensor pads as the sensor circuit board. The sensor circuit board was cured using a reflow oven to obtain a cured circuit board; Obtain the gold wire bonding process parameter set, perform gold wire bonding operation based on the gold wire bonding process parameter set and the solidified circuit board, and obtain the integrated circuit board.

3. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 2, characterized in that, The process of obtaining the gold wire bonding process parameter set includes: A knowledge graph of gold wire bonding is constructed based on pre-built historical data of gold wire bonding; Extract multiple process parameters, their value ranges, and constraint rules from the knowledge graph; Extract a process parameter value range from multiple process parameter value ranges sequentially to obtain the target value range, and then perform the following operations on the target value range: Based on the preset number of process parameters, the target value range is divided into intervals to obtain multiple value intervals. One value interval is extracted from the multiple value intervals in turn, and a numerical extraction operation is performed on the extracted value interval to obtain the process parameter values. The process parameter values ​​are summarized to obtain multiple process parameter values ​​corresponding to the target value range. The multiple process parameter values ​​are recorded as the target process scattered value group. Summarize the target process scattered value groups to obtain multiple target process scattered value groups corresponding to multiple process parameter value ranges. Perform parameter classification operation based on element-distinctness on multiple target process scattered value groups based on multiple process parameters to obtain multiple initial process parameter sets. The initial process parameter sets include multiple initial process parameters, and the initial process parameters correspond one-to-one with the process parameters. If multiple initial process parameter sets all meet the constraint rules, then the multiple initial process parameter sets are confirmed as multiple process parameter sets. The gold wire bonding process parameter set is obtained based on multiple process parameter sets.

4. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 3, characterized in that, The process of obtaining the gold wire bonding process parameter set based on multiple process parameter sets includes: Multiple candidate process parameter sets are generated based on multiple process parameter sets; Simulation operations are performed on multiple candidate process parameter sets to obtain multiple simulated integrated circuit boards, wherein the simulated integrated circuit boards and candidate process parameter sets correspond one-to-one; A board splitting operation is performed on multiple simulated integrated circuit boards to obtain multiple simulated testing component sets. Each simulated testing component set includes multiple simulated testing components, and each simulated testing component set corresponds one-to-one with a simulated integrated circuit board. Performance prediction is performed on multiple simulation detection component sets to obtain multiple performance prediction sets. Each performance prediction set includes multiple performance prediction groups. There is a one-to-one correspondence between the simulation detection component sets and the performance prediction sets. Each performance prediction group includes multiple performance prediction values, and there is a one-to-one correspondence between the performance prediction groups and the simulation detection components. Multiple performance qualities are calculated based on multiple performance prediction sets and pre-built performance quality calculation formulas. The multiple performance qualities are compared with the pre-built performance quality threshold. If there is a performance quality among the multiple performance qualities that is greater than or equal to the performance quality threshold, the performance quality with the largest value among the multiple performance qualities is extracted to obtain the maximum performance quality. The candidate process parameter set corresponding to the maximum performance quality is determined from the multiple candidate process parameter sets to obtain the gold wire bonding process parameter set; If multiple performance quality values ​​are all less than the performance quality threshold, then multiple candidate process parameter sets are used as the multiple process parameter sets, and the step of generating multiple candidate process parameter sets based on multiple process parameter sets is returned until the gold wire bonding process parameter set is obtained.

5. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 4, characterized in that, The calculation of multiple performance qualities based on multiple performance prediction sets and pre-built performance quality calculation formulas includes: Extract a performance prediction set from multiple performance prediction sets sequentially to obtain the target performance prediction set, and then perform the following operations on the target performance prediction set: For each performance prediction group in the target performance prediction group set, a vector construction operation is performed to obtain multiple prediction performance vectors, and a prediction performance matrix is ​​generated based on the multiple prediction performance vectors. Multiple performance types are identified based on the predicted performance matrix, multiple weights are assigned to the multiple performance types, and a weight vector is constructed using the multiple weights. Calculate the weight performance matrix based on the weight vector and the prediction performance matrix; Normalize the weight performance matrix to obtain the normalized performance matrix, as shown below: ， in, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the target performance prediction group set The first performance prediction group A normalized performance value, Indicates the first in the performance prediction group An index of normalized performance values, Indicates the first in the target performance prediction group set Indexes for performance prediction groups; By summing the normalized performance matrices, multiple normalized performance matrices are obtained; Multiple performance qualities are calculated based on multiple normalized performance matrices and performance quality calculation formulas.

6. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 5, characterized in that, The calculation of multiple performance qualities based on multiple normalized performance matrices and performance quality calculation formulas includes: Extract one of the multiple performance types sequentially to obtain the target performance type, and perform the following operations on the target performance type: Extract multiple normalized performance values ​​corresponding to the target performance type from the multiple normalized performance matrices, and identify the normalized performance value with the largest value and the normalized performance value with the smallest value among the multiple normalized performance values ​​to obtain the maximum performance value and the minimum performance value. The maximum and minimum performance values ​​are summarized separately to obtain the maximum performance value vector and minimum performance value vector corresponding to multiple performance types. The maximum performance value vector includes multiple maximum value elements, and each maximum value element corresponds one-to-one with the performance type. Extract one of the multiple normalized performance matrices sequentially to obtain the target normalized performance matrix. Perform the following operations on the target normalized performance matrix: Based on the multiple simulation detection component sets, multiple normalized performance vectors are obtained from the target normalized performance matrix. Then, one normalized performance vector is extracted from the multiple normalized performance vectors to obtain the target performance vector. The following operations are performed on the target performance vector: Calculate the Euclidean distance between the target performance vector and the maximum performance value vector to obtain the positive ideal distance; calculate the Euclidean distance between the target performance vector and the minimum performance value vector to obtain the negative ideal distance. Performance similarity is calculated based on positive and negative ideal distances; Summarize the performance similarity scores to obtain multiple performance similarity scores, and construct the performance similarity vector corresponding to the target normalized performance matrix based on the performance similarity scores; By summing the performance proximity vectors, multiple performance proximity vectors corresponding to multiple performance prediction sets are obtained, where there is a one-to-one correspondence between the performance prediction sets and the performance proximity vectors. Multiple performance qualities are calculated based on multiple performance proximity vectors and performance quality calculation formulas.

7. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 6, characterized in that, The calculation of multiple performance qualities based on multiple performance proximity vectors and performance quality calculation formulas includes: Extract one of the multiple normalized performance matrices sequentially to obtain the key normalized performance matrix. Perform the following operations on the key normalized performance matrix: Based on the multiple performance types, multiple key performance vectors are obtained from the key normalized performance matrix. Each of the multiple key performance vectors is the transpose of the column vector of the key normalized performance matrix, and the key performance vectors correspond one-to-one with the performance types. Each key performance vector includes multiple key performance values. Extract one key performance vector from multiple key performance vectors sequentially to obtain the target key performance vector, and then perform the following operations on the target key performance vector: Calculate the standard deviation of multiple key performance values ​​to obtain the performance standard deviation; Summarize the performance standard deviations to obtain multiple performance standard deviations corresponding to multiple performance types, and construct a performance standard deviation vector based on the multiple performance standard deviations; The performance proximity vector corresponding to the normalized performance matrix represented by the performance standard deviation vector is extracted from multiple performance proximity vectors to obtain a definite proximity vector. The performance quality is then calculated based on the performance standard deviation vector, the definite proximity vector, and the performance quality calculation formula. By summarizing the performance qualities, multiple performance qualities are obtained, each of which corresponds one-to-one with a set of candidate process parameters.

8. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 7, characterized in that, The performance quality calculation formula is as follows: ， in, Indicates performance quality. Indicates the weight of performance similarity. Indicates the standard deviation weight of performance. Indices representing multiple performance similarity values ​​in the performance similarity vector. This represents the first of several proximity values ​​in the performance proximity vector. A proximity index, The first element in the performance proximity vector represents the... Performance similarity Indices representing multiple performance standard deviations of the performance standard deviation vector. The first of several performance standard deviations in the performance standard deviation vector represents the first of the several performance standard deviations. An index with a performance standard deviation, The first element in the performance standard deviation vector represents the... One performance standard deviation.

9. The automated manufacturing method for the integrated temperature and liquid level detection probe as described in claim 8, characterized in that, The process of performing gold wire bonding based on the gold wire bonding process parameter set and the cured circuit board to obtain an integrated circuit board includes: The gold wire bonding process parameter set is imported into the pre-built gold wire bonding machine to obtain the setting bonding machine. The setting bonding machine is used to confirm the positions of multiple sensor groups and multiple bonding position groups in the cured circuit board. Among them, the bonding position group includes a temperature bonding position and a liquid level bonding position. The location of a sensor group is determined sequentially from multiple sensor group locations to obtain the target sensor group location. The following operations are then performed on the target sensor group location: Temperature metal wire and liquid level metal wire are obtained, and a discharge operation is performed on the temperature metal wire and liquid level metal wire using a setting bonding machine to obtain temperature metal ball and liquid level metal ball. The temperature metal ball and the liquid level metal ball are pressed together to the target sensor group position to obtain the temperature welding metal wire and the liquid level welding metal wire; Determine the target binding position group corresponding to the target sensor group position among multiple binding position groups to obtain the target binding position group; The temperature welding metal wire and the liquid level welding metal wire are welded to the target bonding position group to obtain the gold wire bonding pad. The gold wire bonding pads are aggregated to obtain multiple gold wire bonding pads, which are then referred to as the integrated circuit board.

10. An automated manufacturing system for an integrated temperature and liquid level detection probe, characterized in that, The system includes: The instruction receiving module is used to receive the detection probe preparation instruction and confirm the preparation environment based on the detection probe preparation instruction. The preparation environment includes multiple sensor groups, printed circuit boards, multiple probe housings and multiple signal cables. The sensor group includes a temperature sensor and a liquid level sensor. The integrated board separation module is used to perform integration operations on multiple sensor groups and printed circuit boards to obtain an integrated circuit board. Adhesive is applied to the integrated circuit board to obtain an adhesive circuit board. Based on multiple sensor groups, the adhesive circuit board is separated to obtain multiple detection components. The sensor groups and detection components correspond one-to-one. The encapsulation module is used to sequentially extract one detection component from multiple detection components to obtain a target detection component. The following operations are performed on the target detection component: extracting the probe housing corresponding to the detection component from multiple probe housings to obtain a target probe housing; aligning the target detection component with the target probe housing and applying pressure to obtain a component housing; performing a heat curing operation on the component housing to obtain a cured detection component; identifying the signal cable corresponding to the cured detection component from multiple signal cables to obtain a target signal cable; performing a soldering operation on the cured detection component and the target signal cable to obtain a signal detection component; performing a potting compound injection operation on the signal detection component to obtain a potted detection probe; performing a heat curing operation on the potted detection probe to obtain an initial detection probe; and summing the initial detection probes to obtain multiple initial detection probes. The quality inspection module is used to perform quality control screening operations based on detection performance testing and packaging inspection on multiple initial detection probes to obtain multiple temperature and liquid level detection probes.

Images