A kind of eutectic soldering fastest temperature rise control method and system without overshoot

Abstract

The application discloses a eutectic welding fastest temperature rising control method and system without overshoot, which sets an ideal fastest temperature rising curve, establishes a parameter adjustment controller of the fastest temperature rising control, calculates a control output of the parameter adjustment controller, performs nonlinear modeling on a eutectic heating control system, obtains an edge joint field distribution model, predicts a predicted temperature of the eutectic heating control system, updates a polyform parameter tensor in the parameter adjustment controller according to the predicted temperature and an expected temperature corresponding to the predicted temperature in the ideal fastest temperature rising curve, and updates model parameters in the edge joint field distribution model according to the predicted temperature and an actual temperature output by the eutectic heating control system, until the actual temperature is controlled to an expected value corresponding to the ideal fastest temperature rising curve, so that the technical problem that the prior art cannot perform fastest temperature rising control without overshoot on the eutectic welding temperature is solved, and the fastest temperature rising control without overshoot on the eutectic welding is realized.

Description

Technical Field

[0001] This invention mainly relates to the field of eutectic patch welding technology, specifically a method and system for controlling the fastest temperature rise in eutectic welding without overshoot. Background Technology

[0002] Optical communication technology forms the foundation of human communication networks. Whether communicating via wired or wireless networks, information ultimately needs to be transmitted through optical communication technology and then processed before reaching the destination. Even current fifth-generation wireless communication technology requires extensive optical communication infrastructure to connect base stations and achieve high-speed, high-volume data communication. Therefore, for a long time to come, optical communication technology will remain the fundamental basis for the construction of human communication networks. In optical communication, electrical signals are first converted into optical signals, which are then transmitted through optical fibers for propagation. Once at the target optical communication device, the optical signal is converted into other signals for further propagation.

[0003] To ensure stable optical communication, the quality of optical emitting components must be fully guaranteed. As the production equipment for optical emitting components, the eutectic bonding machine's soldering quality significantly impacts the operational quality of these components. During soldering, the chip is placed in the solder area of ​​the substrate, and then the substrate is heated by a heating stage to melt the solder. After cooling, the substrate and chip are soldered together. When heating the substrate, it is necessary to ensure that the temperature of the solder area rises rapidly to the set soldering temperature to melt the solder evenly and ensure the chip can adhere to the substrate; simultaneously, the temperature must not be too high to avoid affecting the soldering quality or damaging the components. Therefore, the soldering temperature rise profile needs to be as rapid as possible without overshoot.

[0004] Currently, there are very few methods for temperature control in eutectic bonding machines. There are two main types of welding temperature control methods in other fields: one is temperature control based on experience, and the other is control of the welding process based on control principles.

[0005] For the first type of method, the experience comes from the control experience in historical production. The subsequent controller parameters and control output are given with reference to these experiences. This type of method is often open-loop control, which is relatively easy to implement, but it cannot achieve fine control and is relatively weak in the face of disturbances in the system.

[0006] The second type of method incorporates closed-loop control and feedforward control, typically using a PID controller or an improved version thereof. This type of method generally requires obtaining a mathematical model between the input and output of the controlled object to achieve higher control accuracy. However, when the system model is difficult to obtain, this type of method becomes limited and may even have negative consequences.

[0007] Chinese patent application CN114063668B discloses a method and system for intelligent temperature control in the production of miniLED circuit boards. This system trains a neural network model using a welding relationship database. The trained network model, combined with welding information acquired during welding, generates a temperature control strategy to control the welding temperature. The system can also update and optimize the model during operation. This method integrates the relationship between welding strategy and welding information into the neural network model, achieving intelligent optimization of the temperature control strategy after inputting welding information. Although this method associates welding information and control strategy through a neural network, its essence is experience transfer, which is not suitable for the fine control required in eutectic welding.

[0008] Chinese patent application CN112404713B discloses an OLED laser welding system and temperature control method. This system includes a laser, a welding head, and a temperature detection module. It detects the temperature of the weld joint during the welding process, obtaining precise temperature information and feeding it back to the control module. The control module then determines the relationship between the welding temperature and the desired temperature; if the welding temperature is higher than the desired temperature, the power is reduced, and vice versa. This temperature setting device and method allows for precise acquisition of the weld joint temperature, which is then used to adjust the control. However, although it incorporates weld joint temperature feedback, the adjustment method is too coarse, resulting in poor actual temperature setting and an inability to quickly and accurately track the ideal temperature.

[0009] Application CN108873985A discloses a temperature control method and system for laser soldering. The system's soldering station includes a temperature sensor for measuring the solder joint temperature and a controller. During operation, the system uses the temperature sensor on the heating station to measure the soldering temperature and feeds it back to the controller. The controller then uses fuzzy PID control to perform feedback control based on the solder joint temperature. Furthermore, the temperature setting method models the mathematical relationship between the solder joint temperature and laser power, and uses this model for feedforward compensation to quickly eliminate errors. This device utilizes fuzzy PID and a mathematical model of solder joint temperature and power for feedforward and feedback soldering temperature control, achieving rapid error elimination. However, the mathematical modeling of solder joint temperature and power in this system and method is relatively simple, and the model is not corrected in real time, making it difficult to reflect the true mathematical relationship between the two. This leads to a decrease in the effectiveness of the feedforward and may even have negative consequences. Therefore, this method is difficult to quickly set the ideal heating curve. Summary of the Invention

[0010] The present invention provides a method and system for controlling the fastest temperature rise in eutectic welding without overshoot, which solves the technical problem that the prior art cannot control the temperature rise of eutectic welding with the fastest temperature rise without overshoot.

[0011] To solve the above-mentioned technical problems, the present invention proposes a method for controlling the fastest temperature rise in eutectic welding without overshoot, comprising:

[0012] Set the ideal fastest heating curve;

[0013] Establish a parameter control controller for the fastest temperature rise control, which includes a first controller and a second controller;

[0014] By combining the first and second controllers, the parameters are calculated to adjust the control output of the controllers.

[0015] Nonlinear modeling of the eutectic heating control system is performed to obtain the edge joint field distribution model;

[0016] Based on the control quantity output by the parameter adjustment controller and the edge joint field distribution model, the predicted temperature of the eutectic heating control system is predicted.

[0017] Update the polymorphic parameter tensor in the parameter control controller based on the predicted temperature and the desired temperature corresponding to the predicted temperature in the ideal fastest heating curve.

[0018] Based on the predicted temperature and the actual temperature output by the eutectic heating control system, the model parameters in the edge joint field distribution model are updated until the actual temperature is controlled to the desired value corresponding to the ideal fastest heating curve, thereby achieving the fastest heating of eutectic welding without overshoot.

[0019] Furthermore, the model of the first controller is as follows:

[0020]

[0021]

[0022] Where, ΔT k (t+1) represents the system output increment at time t+1 during the k-th bonding cycle of eutectic heating control. k (t+1) and T k-1 (t+1) represents the system output at time t+1 for the k-th and (k-1)-th bonding cycles controlled by eutectic heating, respectively. and These are the pseudo-manifold control outputs at time t for the k-th and (k-1)-th patching cycles, respectively. and θ k (t) represents the pseudo-manifold control output increment and pseudo-manifold control quantity at time t during the k-th patching cycle, respectively.

[0023] Furthermore, the second controller is a PID controller.

[0024] Furthermore, by combining the first and second controllers, the formula for calculating the control quantity output by the parameter adjustment controller is as follows:

[0025]

[0026] in, For the k-th patching cycle, the parameter control controller at time t is in the polymorphic parameter tensor The output control quantity under the action, For the k-th patching cycle, the parameter adjustment controller at time t-1 is in the polymorphic parameter tensor The output control quantity under the action, For polymorphic parameter tensors The j-th element in θ k (t) represents the pseudo-manifold of the first controller at time t during the k-th patching cycle, P k (t), I k (t) and D k (t) represents the first, second, and third error adjustment parameters of the second controller in the k-th patching cycle, ε j (t) is the polymorphic parameter action tensor ε(t) = [ΔT] k (t),e k (t),∑e k (t),e k (t)-e k The j-th element in [t-1)], ΔT k (t)=T d (t)-T k-1 (t) represents the system output at time t during the (k-1)th patching cycle and the deviation from the expected value corresponding to the ideal fastest heating curve, e k (t)=T d (t)-T k (t) represents the system deviation at time t during the k-th patching cycle, T k (t) and T d (t) represents the system output at time t during the k-th patching cycle and the expected value corresponding to the ideal fastest heating curve, respectively.

[0027] Furthermore, based on the control quantity output by the parameter adjustment controller and the edge joint field distribution model, the formula for calculating the predicted temperature of the eutectic heating control system is as follows:

[0028]

[0029]

[0030] L i (x)=ω i ·x+b i ,

[0031] in, Let be the predicted temperature at time t for the k-th patching cycle. For the k-th patching cycle, the parameter control controller at time t is in the polymorphic parameter tensor The output control quantity under the action, For the edge joint field distribution model with input as The corresponding model output at that time, Let x be the input to the edge joint field distribution model, n be the number of network layers in the edge joint field distribution model, and σ be the activation function. For linear operators, L i (x) represents the output of the i-th layer network of the edge joint field distribution model, ω i and b i Let be the first and second parameters of the i-th layer network, respectively, and 1 ≤ i ≤ n.

[0032] Furthermore, based on the predicted temperature and the desired temperature corresponding to the predicted temperature in the ideal fastest heating curve, the calculation formula for the polymorphic parameter tensor in the parameter control controller is updated as follows:

[0033]

[0034] in and Let be the polymorphic parameter tensor and the polymorphic parameter tensor increment of the first and second controllers at time t during the k-th patching cycle, respectively. Let η be the polymorphic parameter tensor of the first controller during the (k-1)th patching cycle and the second controller during the kth patching cycle at time t and time t-1, respectively. k (t) represents the update rate, J represents the deviation between the predicted temperature and the expected temperature, and T d (t) represents the predicted temperature and the expected temperature at time t in the k-th patching cycle, respectively. For the k-th patching cycle, the parameter control controller at time t is in the polymorphic parameter tensor The control quantity output under the action.

[0035] The present invention provides a fast temperature rise control system for eutectic welding without overshoot, comprising:

[0036] The present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the overshoot-free eutectic welding fastest temperature rise control method provided by the present invention.

[0037] This invention proposes a method and system for controlling the fastest temperature rise in eutectic welding without overshoot. The method involves setting an ideal fastest temperature rise curve, establishing a parameter-controlled controller for the fastest temperature rise, simultaneously controlling the first and second controllers, calculating the control quantity output by the parameter-controlled controller, performing nonlinear modeling of the eutectic heating control system to obtain an edge joint field distribution model, predicting the predicted temperature of the eutectic heating control system based on the control quantity output by the parameter-controlled controller and the edge joint field distribution model, and updating the polymorphic parameter tensor in the parameter-controlled controller based on the predicted temperature and the expected temperature corresponding to the predicted temperature in the ideal fastest temperature rise curve. The actual temperature output by the system is used to update the model parameters in the edge joint field distribution model until the actual temperature is controlled to the expected value corresponding to the ideal fastest heating curve. This achieves the fastest heating of eutectic bonding without overshoot, solving the technical problem that existing technologies cannot achieve the fastest heating control of eutectic bonding temperature without overshoot. It provides a method for rapid heating control without overshoot for the heating stage of the eutectic bonding machine when the mathematical model of the heating stage-substrate-chip heating bonding system cannot be accurately obtained. By using historical production data to output the control quantity, the rapid heating without overshoot is controlled, ensuring the speed of the bonding heating process and improving the working efficiency of the bonding machine.

[0038] Key aspects of this invention include:

[0039] The purpose of this invention is to design a method for controlling the fastest temperature rise in eutectic bonding without overshoot. This method provides a fast, overshoot-free temperature rise control for the heating stage of a eutectic bonding machine when the mathematical model of the heating and bonding system consisting of the heating stage, substrate, and chip cannot be accurately obtained. By using historical production data to output control quantities, the overshoot-free temperature rise is controlled, ensuring the speed of the bonding heating process and improving the working efficiency of the bonding machine.

[0040] The purpose of this invention is to design a method for the fastest temperature rise control of eutectic bonding without overshoot. This method provides a fastest temperature rise control method without overshoot for the heating station of a eutectic bonding machine, which reduces dynamic errors by updating the control quantity using historical production data. This method enables the solder temperature rise curve to track the desired temperature rise curve, meets process requirements, and avoids the situation of excessive temperature damaging the material, thereby improving the welding quality and yield.

[0041] The beneficial effects of this invention include:

[0042] 1. The key point of this invention is the proposed polymorphic tensor control method. The proposed polymorphic tensor control method can use historical parameter data and current error data in the production process to update and correct the current controller polymorphic parameter tensor, so that the controller can control the solder temperature to the expected temperature at the next moment even when the system model is not accurately obtained, thus achieving the fastest overshoot-free temperature rise control.

[0043] 2. The key point of this invention lies in proposing an edge joint field distribution network model, which fits the complex nonlinear relationship between system input and output. Since the true field distribution model of the system is unavailable, the edge joint field distribution model is proposed to fit the system's input-output relationship. System parameters are adjusted in real time based on input and output data during the production process, improving the characterization capability of the edge joint field distribution model and providing a model basis for the parameter update strategy of the multi-shape controller in the fastest overshoot-free heating control.

[0044] 3. The key point of this invention lies in proposing a real-time update strategy for the controller's polymorphic parameter tensor in a polymorphic control method. The model predicts the output under the current parameter tensor, and the stationary point principle is used to update the polymorphic parameter tensor in real time, ensuring that the updated parameter tensor controls the output temperature to the expected level. After multiple rounds of parameter updates, the controller parameter tensor sequence will tend to stabilize. Under the action of this stable parameter tensor sequence, the system output will follow the set fastest temperature rise curve without overshoot with extremely high precision. Attached Figure Description

[0045] Figure 1 This is a heating schematic diagram of the fastest temperature rise control method for overshoot-free eutectic welding according to Embodiment 2 of the present invention;

[0046] Figure 2 This is the ideal fastest temperature rise curve without overshoot in Embodiment 2 of the present invention;

[0047] Figure 3 This is a control block diagram of the overshoot-free eutectic welding fastest temperature rise control method according to Embodiment 2 of the present invention;

[0048] Figure 4 This is a flowchart of the fastest temperature rise control method for overshoot-free eutectic welding according to Embodiment 2 of the present invention;

[0049] Figure 5 This is a structural block diagram of the overshoot-free eutectic welding fastest temperature rise control system according to an embodiment of the present invention.

[0050] Figure label:

[0051] U1 is the eutectic heating stage; U2 is the substrate; U3 is the substrate solder; 10 is the memory; 20 is the processor. Detailed Implementation

[0052] To facilitate understanding of the present invention, the present invention will be described more fully and in detail below with reference to the accompanying drawings and preferred embodiments, but the scope of protection of the present invention is not limited to the following specific embodiments.

[0053] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings, but the present invention can be implemented in many different ways as defined and covered by the claims.

[0054] Example 1

[0055] The method for controlling the fastest temperature rise in eutectic welding without overshoot provided in Embodiment 1 of the present invention includes:

[0056] Step S101: Set the ideal fastest heating curve.

[0057] Step S102: Establish a parameter adjustment controller for the fastest temperature rise control. The parameter adjustment controller includes a first controller and a second controller.

[0058] Step S103: Combine the first controller and the second controller, calculate the control quantity output by the parameter adjustment controller.

[0059] Step S104: Perform nonlinear modeling on the eutectic heating control system to obtain the edge joint field distribution model.

[0060] Step S105: Based on the control quantity output by the parameter adjustment controller and the edge joint field distribution model, predict the predicted temperature of the eutectic heating control system.

[0061] Step S106: Update the polymorphic parameter tensor in the parameter adjustment controller according to the predicted temperature and the expected temperature corresponding to the predicted temperature in the ideal fastest heating curve.

[0062] Step S107: Based on the predicted temperature and the actual temperature output by the eutectic heating control system, update the model parameters in the edge joint field distribution model until the actual temperature is controlled to the desired value corresponding to the ideal fastest heating curve, thereby achieving the fastest heating of eutectic welding without overshoot.

[0063] The overshoot-free eutectic welding rapid temperature rise control method provided in this invention establishes an ideal rapid temperature rise curve, sets up a parameter regulating controller for rapid temperature rise control, combines a first controller and a second controller, calculates the control quantity output by the parameter regulating controller, performs nonlinear modeling of the eutectic heating control system, obtains an edge joint field distribution model, predicts the predicted temperature of the eutectic heating control system based on the control quantity output by the parameter regulating controller and the edge joint field distribution model, updates the polymorphic parameter tensor in the parameter regulating controller based on the predicted temperature and the expected temperature corresponding to the predicted temperature in the ideal rapid temperature rise curve, and updates the parameter regulating controller based on the predicted temperature and the expected temperature corresponding to the predicted temperature in the eutectic heating control system. The actual temperature output by the system is used to update the model parameters in the edge joint field distribution model until the actual temperature is controlled to the expected value corresponding to the ideal fastest heating curve. This achieves the fastest heating of eutectic bonding without overshoot, solving the technical problem that existing technologies cannot achieve the fastest heating control of eutectic bonding temperature without overshoot. It provides a method for rapid heating control without overshoot for the heating stage of the eutectic bonding machine when the mathematical model of the heating stage-substrate-chip heating bonding system cannot be accurately obtained. By using historical production data to output the control quantity, the rapid heating without overshoot is controlled, ensuring the speed of the bonding heating process and improving the working efficiency of the bonding machine.

[0064] Example 2

[0065] Figure 1 This diagram illustrates the heating and soldering of a substrate and chip using a eutectic heating stage. U1 represents the eutectic heating stage, U2 the substrate, and U3 the substrate solder. At the start of each soldering cycle, the robotic arm places the substrate on the heating stage, then places the chip on the solder area of ​​the substrate. The heating stage then rapidly heats to the set temperature, avoiding large overshoot to prevent chip burnout. After reaching the set temperature, the stage is stabilized at that temperature for a period to ensure proper adhesion between the chip and solder. The stage is then rapidly cooled to complete the soldering. This process is repeated to achieve mass chip soldering.

[0066] Current temperature control methods struggle to achieve rapid, overshoot-free temperature control when the system model is unclear. To balance the efficiency and yield of eutectic bonding machines, this invention provides a fastest, overshoot-free temperature rise control method. This method utilizes extensive historical and real-time data from the production process to adjust controller parameters, ensuring the output curve meets the desired fastest, overshoot-free temperature rise curve.

[0067] Each eutectic bonding cycle shares similar initial conditions, the same controlled object, and the same control objective. Identical control parameters will produce similar control results in different bonding cycles. By inheriting the best parameters and improving the worst ones, a superior set of control parameters can be obtained, ensuring that the control results at each moment in the bonding cycle closely match the desired fastest overshoot-free heating curve. Simultaneously, real-time data is used to suppress unknown disturbances in the control loop, enhancing the controller's robustness.

[0068] To achieve tracking of the ideal fastest temperature rise curve without overshoot, the fastest temperature rise control method proposed in this invention includes the following parts:

[0069] Design as attached Figure 2 The temperature rise curve T shown d (t), T d (t) represents the fastest temperature rise curve without overshoot under the premise of meeting the production process and production conditions.

[0070] Design as attached Figure 3 The polymorphic tensor control method shown precisely tracks the output curve during the production process at T. d (t) achieves the fastest temperature rise control without overshoot.

[0071] The parameter adjustment controller for the polymorphic tensor control method consists of a first controller and a second controller, specifically corresponding to pseudo-manifold control and error adjustment control. The controller operates at time t during the k-th patch cycle on the polymorphic parameter tensor. The control quantity of the output under action As shown in equation (1).

[0072]

[0073] in, For polymorphic parameter tensors The j-th element in θ k (t) represents the pseudomanifold quantity of the pseudomanifold control in the controller at time t during the k-th production process, reflecting the system output increment ΔT. k (t+1)=T k (t+1)-T k-1 (t+1) and pseudo-manifold control output increment The relationship between them is shown in equation (2), P k (t), I k (t) and D k (t) represents the error adjustment parameter of the error adjustment control section in the k-patch cycle, ε j (t) is the polymorphic parameter action tensor ε(t) = [ΔT] k (t),e k(t),∑e k (t),e k (t)-e k The j-th element in [t-1)], ΔT k (t)=T d (t)-T k-1 (t) represents the system output at time t during the (k-1)th patching cycle and the deviation from the expected value corresponding to the ideal fastest heating curve, e k (t)=T d (t)-T k (t) represents the deviation at time t during the k patch period.

[0074]

[0075] pseudo-manifolds have The characteristic of is that c is a constant. This characteristic can be proved as follows:

[0076] For a controlled nonlinear system with pseudo-manifold control as shown in equation (3), its output increment has an increment form as shown in equation (4) here.

[0077] T(t+1)=f(T(t),T(t-1),···,T(tn T ),U f (t),U f (t-1),···,U f (tn Uf ))(3)

[0078]

[0079] Where, n T and n Uf Let f(···) denote the order of the system, and f(···) be a nonlinear function between the input and output.

[0080] By the mean value theorem, we can obtain:

[0081]

[0082] in, This represents the function f(···) with respect to u. k A suitable partial derivative value of (t), ξ k (t) represents the following:

[0083]

[0084] Consider the variable α as shown in equation (7). k The equation for (t).

[0085] ξk (t)=α k (t)Δu k (t) (7)

[0086] When |Δu k When (t)|≠0, the equation obviously has a solution. Let Equation (2) can then be derived from equation (6). Making the assumption shown in equation (8), that is, the system satisfies generalized Lipschitz continuity, it is obvious... Established.

[0087]

[0088] Design a strategy for updating the controller's polymorphic parameter tensor. To update the controller's polymorphic parameter tensor, it is necessary to obtain the predicted output of the control quantity under the current polymorphic parameter tensor, and update the controller parameter tensor based on the deviation between the predicted output and the expected output.

[0089] A model of the eutectic heating control system is constructed. For this system, the input and output are the solutions to the boundary conditions and field equations in the temperature field model, respectively, making accurate modeling and analytical solution extremely difficult. Therefore, an edge joint field distribution network is designed to fit the complex nonlinear relationship between the input and output of the eutectic heating system. The specific process is as follows:

[0090] Edge joint field distribution network It can be represented as:

[0091]

[0092] Where x is the network input and n is the number of network layers. Let L be the weight matrix. m (1≤m≤n) represents the output of each layer, calculated according to equation (10), where σ is the activation function. It is a linear operator.

[0093] L i (x)=ω i ·x+b i (10)

[0094] The parameters of the edge joint field distribution network are updated in real time using the deviation between each prediction result and the actual result.

[0095] Predicting the action of the time parameter tensor ε(t) at time t in the parameter The system output below for:

[0096]

[0097] The desired control parameters should adjust the predicted output to the desired output, such that the deviation between the predicted output and the desired output is minimized. The value approaches 0. The controller parameter tensor is updated using the stationary point principle.

[0098]

[0099]

[0100] Where, η k (t) represents the update rate, which is dynamically selected based on experience. A larger update rate is set in the first few production cycles, and a smaller value is set as the parameters gradually stabilize. Furthermore, within a welding cycle, a larger value is needed during the heating phase, and the value is adjusted to a smaller value after heating to the set temperature. The partial derivatives are calculated as follows.

[0101]

[0102] The control quantity is adjusted and output using the updated controller polymorphic parameter tensor.

[0103] The output control quantity acts on the eutectic heating stage to heat the solder area, and measures the actual temperature T of the solder area after the control quantity is applied. k (t), using predicted temperature With actual temperature T k The deviation between (t) and the system model Parameters in Adjustments were made to make the fitted system model closer to the real model.

[0104] For a single eutectic heating welding process, the controller parameters and system model parameters are gradually optimized to reduce the error. Upon entering the next welding cycle, the controller parameters are initialized to... The system model parameters are initialized to N is the total number of control steps in the previous heating cycle.

[0105] After parameter adjustments over several welding cycles, the controller parameter tensor sequence and system model parameters It will gradually converge to a small neighborhood of a value, and the tracking error will converge to a very small value.

[0106] Example 3

[0107] The implementation method will now be described in detail with reference to the accompanying drawings.

[0108] The method for controlling the fastest temperature rise in eutectic welding without overshoot according to embodiments of the present invention includes:

[0109] Step 1, as shown in the attached document Figure 1 The diagram illustrates the heating and soldering of a substrate and chip using a eutectic heating stage. At the start of each soldering cycle, the robotic arm places the substrate on the heating stage, then places the chip on the solder area of ​​the substrate. The heating stage then rapidly heats to the set temperature, avoiding large overshoot to prevent chip burnout. After reaching the set temperature, the stage is stabilized at that temperature for a period to ensure proper adhesion between the chip and solder. The stage is then rapidly cooled to complete the soldering process. This process is repeated to achieve mass chip soldering.

[0110] Step 2, design as attached Figure 2 The temperature rise curve T shown d (t), T d (t) represents the fastest temperature rise curve without overshoot under the premise of meeting the production process and conditions. The output curve during the production process is precisely tracked at T. d The fastest temperature rise curve without overshoot can be set on (t).

[0111] Step 3, as shown in the attached document Figure 3 The designed polymorphic tensor control method and its attachment Figure 4 The process shown achieves the fastest overshoot-free temperature control for eutectic bonding.

[0112] Step 4: Initialize the polymorphic parameter tensor of polymorphic tensor controller U5 according to equation (15).

[0113]

[0114] The polymorphic parameter tensor is defined as follows: θ k (t) represents the pseudomanifold quantity of the pseudomanifold control in the controller at time t during the k-th production process, reflecting the system output increment ΔT. k (t+1)=T k (t+1)-T k-1 (t+1) and pseudo-manifold control output increment The relationship between them is shown in equation (2), P k (t),、I k (t) and D k (t) represents the error adjustment parameter of the error adjustment control section in the k-th patching cycle.

[0115] Step 5: Initialize the edge joint field distribution model parameters according to equation (16). Obtain the initial edge joint field distribution network model

[0116]

[0117] Where N is the number of control steps in the previous welding cycle. Here, x is a randomly initialized set of model parameters, x is the network input, and n is the number of network layers. Let L be the weight matrix. m (1≤m≤n) represents the output of each layer, and σ is the activation function. It is a linear operator.

[0118] Step 6, calculate the time t of the controller in the polymorphic parameter tensor during the k-th patch cycle according to equation (1). The control quantity of the output under action

[0119] Step 7, output the control signal. The input is fed into the edge joint field distribution model, and the predicted output is obtained according to Equation (11).

[0120] Step 8: Calculate each derivative according to equation (14) to update the controller parameters.

[0121] Step 9: Adjust and update the controller parameters in equations (12) and (13) using the calculated derivatives.

[0122] Step 10: Calculate the controller output using the updated controller parameter tensor.

[0123]

[0124] Step 11, output the control quantity at time t In a system that operates on a eutectic heating stage and a substrate chip, the actual temperature T after the application of a control variable is measured. k (t).

[0125] Step 12: Adjust the system model parameters using the deviation between the actual temperature and the predicted temperature.

[0126] Step 13: After completing one step of control, return to Step 6 for the next step of control. Repeat the optimization and control until the current heating and welding cycle ends, then return to Step 4 to control the next heating and welding cycle.

[0127] Reference Figure 5 The overshoot-free eutectic welding fastest temperature rise control system proposed in this embodiment includes a memory 10, a processor 20, and a computer program stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program, it implements the steps of the overshoot-free eutectic welding fastest temperature rise control method proposed in this embodiment.

[0128] The specific working process and working principle of the overshoot-free eutectic welding fastest temperature rise control system in this embodiment can be referred to the working process and working principle of the overshoot-free eutectic welding fastest temperature rise control method in this embodiment.

[0129] The above are merely preferred embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A eutectic soldering steepest temperature rising control method without overshoot, characterized by, The method comprises: Setting an ideal fastest temperature rising curve; Establishing a parameter adjustment controller of the fastest temperature rising control, the parameter adjustment controller comprising a first controller and a second controller, wherein a model of the first controller is: ， ， in, For eutectic heating control The next patch cycle is in The system output increment at time t. and These are the first eutectic heating control Second and third The next patch cycle is in The system output at any given time, and The first Second and third The next patch cycle is in The pseudo-manifold control output at time t, and The first The next patch cycle is in The pseudo-manifold control output increment and pseudo-manifold control quantity at time t, the second controller is a PID controller; Solving the first controller and the second controller together to calculate a control amount output by the parameter adjustment controller, wherein a calculation formula for calculating the control amount output by the parameter adjustment controller when the first controller and the second controller are solved together is: ， wherein, is the th patch period, the parameter adjustment controller outputs a control quantity at time under the action of the multiphase parameter tensor , is the th patch period, the parameter adjustment controller outputs a control quantity at time under the action of the multiphase parameter tensor , is the th element in the multiphase parameter tensor , is the pseudo-manifold quantity of the first controller at time , and are respectively the first, second and third error adjustment parameters of the second controller in the th patch period, is the th element in the multiphase parameter action tensor , is the th patch period the deviation between the system output at time and the expected value corresponding to the ideal fastest temperature rising curve, is the system deviation at time in the th patch period, are respectively the system output at time in the th patch period and the expected value corresponding to the ideal fastest temperature rising curve. Nonlinear modeling of the eutectic heating control system to obtain an edge joint field distribution model; According to the control amount output by the parameter adjustment controller and the edge joint field distribution model, predicting a predicted temperature of the eutectic heating control system; According to the predicted temperature and an expected temperature corresponding to the predicted temperature in the ideal fastest temperature rising curve, updating a poly-morphic parameter tensor in the parameter adjustment controller; According to the predicted temperature and an actual temperature output by the eutectic heating control system, updating a model parameter in the edge joint field distribution model until the actual temperature is controlled to an expected value corresponding to the ideal fastest temperature rising curve, thereby realizing the fastest temperature rising of the eutectic welding without overshoot.

2. The no overshoot eutectic soldering snap-heating control method according to claim 1, characterized by, According to the control amount output by the parameter adjustment controller and the edge joint field distribution model, a calculation formula for predicting the predicted temperature of the eutectic heating control system is: ， ， ， in, For the first The next patch cycle is in Predicted temperature at any time For the first Next placement cycle, parameter adjustment controller At any moment in the polymorphic parameter tensor The output control quantity under the action, For the edge joint field distribution model with input as The corresponding model output at that time, For the edge joint field distribution model, As input to the edge joint field distribution model, The number of network layers in the edge joint field distribution model. For activation function, For linear operators, The first edge joint field distribution model The output of the layer network, and The first The first and second parameters of the layer network, and .

3. The no overshoot eutectic soldering snap-heating control method according to claim 2, characterized by According to the predicted temperature and an expected temperature corresponding to the predicted temperature in the ideal fastest temperature rising curve, a calculation formula for updating the poly-morphic parameter tensor in the parameter adjustment controller is: in and The first controller and the second controller are respectively The next patch cycle is in The polymorphic parameter tensor and the polymorphic parameter tensor increment at time t. For the first controller Next placement cycle and second controller The next patch cycle is respectively in Time and The polymorphic parameter tensor at time t, For update rate, To account for the deviation between the predicted temperature and the expected temperature, and , and The first The next patch cycle is in Predicted and expected temperatures at any given time. For the first Next placement cycle, parameter adjustment controller At any moment in the polymorphic parameter tensor The control quantity output under the action.

4. A fastest temperature rising control system of eutectic welding without overshoot, the system comprising: A memory (10), a processor (20), and a computer program stored in the memory (10) and executable on the processor (20), characterized in that the processor (20) implements the steps of the method of any one of claims 1 to 3 when executing the computer program.

Images