A temperature-pressure double feedback-based locking force self-adaptive control method

Abstract

The present application relates to a temperature-pressure double feedback based locking force self-adaptive control method, which collects temperature field data in real time through a temperature sensor array, online identifies the current equivalent elastic modulus and equivalent damping coefficient of the intelligent buffer material based on a temperature-mechanical property mapping database, and uses them for dynamic compensation correction of the pressure field data to generate an effective pressure distribution estimation value on the chip surface, solving the control lag problem caused by the mutual fragmentation of temperature control and force control and the inability to perceive the real-time mechanical state changes of the material. Based on the compensated effective pressure distribution, the chip safety pressure-bearing state is evaluated, preventive fine-tuning instructions are generated in the early warning state, the transition from passive emergency braking to active predictive protection is realized, the control lag and regulation overshoot caused by the thermal-force coupling effect are eliminated, the response speed and regulation accuracy of the locking force control are significantly improved, the occurrence probability of chip cracking and edge collapse is reduced, and the packaging yield is improved.

Description

Technical Field

[0001] This invention relates to the field of temperature and pressure locking control technology, and specifically to a locking force adaptive control method based on temperature-pressure dual feedback. Background Technology

[0002] In advanced semiconductor packaging processes, mounting and latching bare die chips is a critical step. The latching head applies vertical pressure to press the chip onto the substrate. The magnitude and uniformity of the latching force directly determine the chip's survival rate and packaging yield.

[0003] To protect highly brittle bare die chips from pressure damage, the industry commonly places a layer of temperature-responsive smart buffer material between the latching head and the chip. This material's elastic modulus and damping characteristics change significantly with temperature. The force control of the latching device primarily employs a single-variable closed-loop control scheme, using the contact pressure collected by a pressure sensor as the sole feedback quantity. The temperature control loop and the force control loop operate independently, lacking information exchange and collaborative adjustment mechanisms. However, during the latching process, the temperature field distribution of the smart buffer material is constantly changing due to frictional heat generation, ambient temperature fluctuations, and heat conduction at the contact interface. The material's equivalent elastic modulus can shift significantly with temperature changes, altering the dynamic response characteristics of the pressure transmission link. Existing single-pressure feedback schemes cannot detect real-time changes in the material's mechanical state. When the pressure sensor detects an abnormal pressure value, the material's mechanical state has already substantially changed due to temperature variations. Feedback adjustments at this point inherently have a time lag, easily leading to overshoot or undershoot, thus failing to effectively protect the chip. Summary of the Invention

[0004] In view of the above problems, the present invention provides an adaptive control method for locking force based on temperature-pressure dual feedback.

[0005] To achieve the above objectives, the present invention provides an adaptive control method for locking force based on temperature-pressure dual feedback, comprising:

[0006] Obtain the latching process parameters, including the target pressure value, the target displacement value, and the equivalent stress safety threshold of the monitoring micro-element on the chip surface;

[0007] The motion speed curve of the locking head is set according to the locking process parameters, the locking process is started, and the temperature field data and pressure field data of the intelligent buffer material are collected synchronously through the temperature sensor array and pressure sensor array, and the displacement data of the locking head is obtained synchronously.

[0008] Based on temperature field data, the temperature-mechanical property mapping database is queried to identify the current equivalent elastic modulus and equivalent damping coefficient of the smart buffer material;

[0009] Based on the current equivalent elastic modulus, equivalent damping coefficient and displacement data, the pressure field data is dynamically compensated and corrected to generate an estimated value of the effective pressure distribution on the chip surface.

[0010] Based on the estimated effective pressure distribution on the chip surface and combined with the equivalent stress safety threshold, the chip's safe pressure-bearing state is assessed, and a global safety risk level label is generated.

[0011] When the global security risk level is marked as a warning state, a preventive fine-tuning instruction is generated. The preventive fine-tuning instruction includes the target adjustment displacement and target adjustment speed of the locking head.

[0012] Execute preventive fine-tuning instructions and repeat the acquisition, identification, compensation correction and evaluation steps until the chip's safe pressure-bearing state meets the preset conditions;

[0013] When the chip's safe pressure-bearing state meets the preset conditions, the locking is determined to be complete and the locking result is output.

[0014] In some embodiments, temperature field data and pressure field data of the smart buffer material are simultaneously acquired through a temperature sensor array and a pressure sensor array, and displacement data of the locking head is simultaneously obtained, including:

[0015] Based on the locking process parameters, a three-stage motion speed curve for the locking head is set, which includes a rapid approach stage, a buffer contact stage, and a pressure loading stage.

[0016] During the movement of the locking head according to the three-segment motion speed curve, the temperature signal output by the temperature sensor array, the pressure signal output by the pressure sensor array, and the displacement signal output by the displacement sensor are collected synchronously at a preset sampling period. The preset sampling period is less than one-fifth of the dominant time constant of the locking head mechanical system.

[0017] The validity of the synchronously acquired temperature, pressure and displacement signals is verified. The validity verification includes temperature range verification, pressure monotonicity verification and displacement-pressure consistency verification.

[0018] Abnormal data that fails the validity check is removed, and the remaining valid data is stamped with a unified timestamp to generate temperature field data, pressure field data, and displacement data.

[0019] In some embodiments, based on temperature field data, a temperature-mechanical property mapping database is queried to identify the current equivalent elastic modulus and equivalent damping coefficient of the smart buffer material, including:

[0020] Spatial interpolation is performed on the discrete temperature values ​​of each temperature measurement point in the temperature field data to reconstruct the continuous two-dimensional temperature field distribution at the contact interface between the smart buffer material and the chip.

[0021] Based on the continuous two-dimensional temperature field distribution, the area-weighted average temperature value of the contact interface is calculated using the area of ​​the influence region represented by each temperature measuring point as the weight, and is used as the characteristic temperature at the current moment.

[0022] Using the characteristic temperature as the query index, table lookup and interpolation are performed in the temperature-mechanical property mapping database to obtain the global equivalent elastic modulus and the global equivalent damping coefficient.

[0023] Calculate the deviation between the temperature value of each temperature measuring point and the characteristic temperature. For measuring points whose deviation exceeds the preset gradient threshold, perform independent lookup interpolation based on the local temperature of that measuring point to obtain the local mechanical state correction matrix.

[0024] In some embodiments, the deviation between the temperature value of each temperature measuring point and the characteristic temperature is calculated. For measuring points where the deviation exceeds a preset gradient threshold, a lookup table interpolation is performed independently using the local temperature of that measuring point to obtain a local mechanical state correction matrix, including:

[0025] Calculate the temperature deviation between the temperature value of each temperature measuring point and the characteristic temperature, and mark the temperature measuring points whose absolute value of the temperature deviation exceeds the preset gradient threshold as local correction measuring points.

[0026] Using the local temperature of each local correction measurement point as a substitute for the characteristic temperature as a query index, table lookup and interpolation are performed in the temperature-mechanical property mapping database to obtain the local equivalent elastic modulus and local equivalent damping coefficient corresponding to each local correction measurement point.

[0027] Based on the spatial location of each local correction measuring point on the contact interface, the local equivalent elastic modulus and local equivalent damping coefficient are organized into a local mechanical state correction matrix. The elements at the non-local correction measuring point locations in the local mechanical state correction matrix are taken as the global equivalent elastic modulus and global equivalent damping coefficient.

[0028] In some embodiments, based on the current equivalent elastic modulus, equivalent damping coefficient, and displacement data, dynamic compensation and correction are performed on the pressure field data to generate an estimate of the effective pressure distribution on the chip surface, including:

[0029] The force transmission process of the smart buffer material is equivalent to a first-order inertial element, and the time constant of the first-order inertial element is determined by the ratio of the current equivalent damping coefficient to the current equivalent elastic modulus.

[0030] For the pressure value sequence of each pressure measuring point in the pressure field data, dynamic compensation calculation is performed using the inverse model of the first-order inertial element. The dynamic compensation calculation is performed in the discrete time domain. The compensated pressure value at the current moment is determined by the original pressure value at the current moment, the compensated pressure value at the previous moment, and the current equivalent elastic modulus and the current equivalent damping coefficient.

[0031] For the pressure measurement points corresponding to the spatial locations of each local correction measurement point in the local mechanical state correction matrix, the local equivalent elastic modulus and local equivalent damping coefficient corresponding to the local correction measurement point are used to replace the current equivalent elastic modulus and current equivalent damping coefficient, and dynamic compensation calculation is performed independently.

[0032] The compensated pressure values ​​of each pressure measuring point are arranged according to their spatial location to generate an estimated value of the effective pressure distribution on the chip surface.

[0033] In some embodiments, the pressure value sequence of each pressure measuring point in the pressure field data is dynamically compensated using the inverse model of a first-order inertial element, including:

[0034] The inverse model of the first-order inertial element is discretized in the discrete time domain to obtain the compensation operation rules in the form of first-order difference equations.

[0035] Based on the current equivalent elastic modulus and the current equivalent damping coefficient, the compensation coefficient of the first-order difference equation is updated in real time. The compensation coefficient is inversely proportional to the time constant of the first-order inertial element.

[0036] For each pressure measuring point, the compensation calculation rules are executed on a sampling cycle basis:

[0037] The compensated pressure value of the current sampling period is equal to the weighted combination of the original pressure value of the current sampling period and the compensated pressure value of the previous sampling period. The weight of the weighted combination is determined by the compensation coefficient updated in real time.

[0038] When the current equivalent elastic modulus increases, the compensation coefficient increases accordingly, which increases the weight of the original pressure value in the current sampling period in the weighted combination.

[0039] When the current equivalent damping coefficient increases, the compensation coefficient decreases accordingly, thereby increasing the weight of the compensated pressure value from the previous sampling period in the weighted combination.

[0040] In some embodiments, based on the estimated effective pressure distribution on the chip surface and combined with the equivalent stress safety threshold, the chip's safe pressure-bearing state is assessed, and a global safety risk level identifier is generated, including:

[0041] The estimated effective pressure distribution on the chip surface is mapped to each monitoring micro-element on the chip surface. Spatial interpolation is performed on the monitoring micro-element that is not directly covered by the pressure measurement point to generate a continuous effective pressure field distribution covering the entire chip surface.

[0042] Calculate the area-weighted average effective pressure value of the continuous effective pressure field distribution, compare the area-weighted average effective pressure value with the compressive strength threshold, and generate a uniform compressive stress assessment result;

[0043] Based on the spatial gradient of the continuous effective pressure field distribution, combined with the chip thickness and back support conditions, the bending stress of each monitoring micro-element is calculated according to the thin plate bending theory. The bending stress is compared with the equivalent stress safety threshold to generate the bending stress assessment result.

[0044] Based on the results of the uniform compressive stress assessment and the bending stress assessment, the overall safety risk level is determined. The overall safety risk level includes three levels: safe state, warning state, and dangerous state.

[0045] In some embodiments, based on the spatial gradient of the continuous effective pressure field distribution, combined with the chip thickness and back support conditions, the bending stress of each monitoring micro-element is calculated according to the thin-plate bending theory, including:

[0046] For each monitoring micro-element, the effective pressure values ​​of the two monitoring micro-elements adjacent to the monitoring micro-element are taken along the first horizontal direction, and the first pressure gradient of the monitoring micro-element in the first horizontal direction is calculated by the central difference method.

[0047] The effective pressure values ​​of two monitoring micro-elements adjacent to the monitoring micro-element are taken along the second horizontal direction. The second pressure gradient of the monitoring micro-element in the second horizontal direction is calculated by the central difference method. The second horizontal direction is orthogonal to the first horizontal direction.

[0048] Based on the first pressure gradient and the second pressure gradient, combined with the chip thickness and the back support conditions, the first bending stress component and the second bending stress component in the second horizontal direction of the monitoring micro-element are calculated according to the thin plate bending theory.

[0049] The bending stress value of the monitoring micro-element is obtained by vector synthesis of the first bending stress component and the second bending stress component.

[0050] In some embodiments, when the global security risk level is identified as a warning state, a preventative fine-tuning instruction is generated. This preventative fine-tuning instruction includes the target adjustment displacement and target adjustment speed of the locking head, and includes:

[0051] The difference between the estimated effective pressure distribution on the chip surface and the target pressure distribution is calculated and used as the desired pressure adjustment amount.

[0052] Based on the inverse solution calculation of the multiphysics coupling calibration model of the locking system, combined with the current equivalent elastic modulus, the target adjustment displacement of the locking head is estimated according to the expected pressure adjustment amount.

[0053] Based on the target adjustment displacement, an S-shaped velocity planning curve with continuously varying acceleration is used to generate the target adjustment speed of the locking head;

[0054] Based on the rate of change of the current chip stress state, the response time limit is adaptively set to generate a preventive fine-tuning instruction that includes the target adjustment displacement, the target adjustment speed, and the response time limit.

[0055] In some embodiments, the movement speed curve of the locking head is set according to the locking process parameters, and the locking process is started, including:

[0056] Based on the specification identification code of the chip to be mounted, retrieve the corresponding latching process parameters from the process database;

[0057] During the rapid approach phase, the locking head is controlled to move toward the chip mounting surface at a first speed until the gap between the end face of the locking head and the upper surface of the smart buffer material is reduced to a preset approach gap threshold.

[0058] When the gap decreases to a preset gap threshold, switch to the buffer contact section and control the locking head to continue descending at a second speed lower than the first speed until the pressure sensor array detects a pressure signal greater than the preset contact determination threshold.

[0059] When the pressure sensor array detects a pressure signal greater than the preset contact determination threshold, it switches to the pressure loading stage and controls the locking head to continue pressing down at the preset loading rate until the target pressure value or target displacement value is reached, and the locking process is started.

[0060] Unlike existing technologies, the above-mentioned solution acquires temperature field data in real time through a temperature sensor array. Based on a temperature-mechanical property mapping database, it identifies the current equivalent elastic modulus and equivalent damping coefficient of the intelligent buffer material online and uses them for dynamic compensation and correction of pressure field data to generate an estimate of the effective pressure distribution on the chip surface. This solves the control lag problem caused by the disconnect between temperature control and force control in existing technologies and the inability to perceive real-time changes in the material's mechanical state. Simultaneously, it assesses the chip's safe pressure-bearing state based on the compensated effective pressure distribution and generates preventative fine-tuning commands in the early warning state. This achieves a shift from passive emergency braking to proactive predictive protection, eliminating control lag and overshoot caused by the thermo-mechanical coupling effect. It significantly improves the response speed and adjustment accuracy of the latching force control, reduces the probability of chip microcracks and edge breakage, and improves packaging yield.

[0061] The above description of the invention is merely an overview of the technical solution of this application. In order to enable those skilled in the art to better understand the technical solution of this application and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of this application easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of this application. Attached Figure Description

[0062] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of the present invention and other related contents, and should not be considered as limitations on this application.

[0063] In the accompanying drawings of the instruction manual:

[0064] Figure 1 This is a schematic diagram illustrating steps S101 to S108 as described in the specific implementation method;

[0065] Figure 2 This is a schematic diagram illustrating steps S201 to S204 as described in the specific implementation method;

[0066] Figure 3 The following is a schematic diagram illustrating steps S301 to S304 as described in the specific implementation method. Detailed Implementation

[0067] To illustrate the possible application scenarios, technical principles, implementable specific solutions, and achievable objectives and effects of this application in detail, the following description, in conjunction with the listed specific embodiments and accompanying drawings, provides a detailed explanation. The embodiments described herein are merely illustrative of the technical solutions of this application and are therefore intended to limit the scope of protection of this application.

[0068] Please see Figure 1 This embodiment provides a locking force adaptive control method based on temperature-pressure dual feedback, including:

[0069] S101. Obtain the latching process parameters, including the target pressure value, the target displacement value, and the equivalent stress safety threshold of the monitoring micro-element on the chip surface;

[0070] S102. Set the motion speed curve of the locking head according to the locking process parameters, start the locking process, and synchronously collect the temperature field data and pressure field data of the intelligent buffer material through the temperature sensor array and pressure sensor array, and synchronously obtain the displacement data of the locking head.

[0071] S103. Based on temperature field data, query the temperature-mechanical property mapping database to identify the current equivalent elastic modulus and equivalent damping coefficient of the smart buffer material.

[0072] S104. Based on the current equivalent elastic modulus, equivalent damping coefficient and displacement data, the pressure field data is dynamically compensated and corrected to generate an estimated value of the effective pressure distribution on the chip surface.

[0073] S105. Based on the estimated effective pressure distribution on the chip surface and combined with the equivalent stress safety threshold, assess the chip's safe pressure-bearing state and generate a global safety risk level label.

[0074] S106. When the global safety risk level is marked as a warning state, a preventive fine-tuning instruction is generated. The preventive fine-tuning instruction includes the target adjustment displacement and target adjustment speed of the locking head.

[0075] S107. Execute the preventive fine-tuning command and repeat the acquisition, identification, compensation correction and evaluation steps until the chip's safe pressure-bearing state meets the preset conditions.

[0076] S108. When the chip's safe pressure-bearing state meets the preset conditions, the locking is determined to be complete and the locking result is output.

[0077] In step S101, the mounting process parameters are retrieved from the process database by the upper control system based on the specification identification code of the chip to be mounted. The target pressure value is used to define the expected force applied by the mounting head to the surface of the smart buffer material. The target displacement value is used to define the expected stroke of the mounting head relative to the initial zero position. The equivalent stress safety threshold is used to define the maximum equivalent stress that each monitoring micro-element on the chip surface can withstand. Each monitoring micro-element is a discrete region divided along the horizontal direction on the chip surface.

[0078] In step S102, the motion speed curve includes a rapid approach segment, a buffer contact segment, and a pressure loading segment. The rapid approach segment drives the locking head to approach the chip mounting surface at a higher speed. The buffer contact segment drives the locking head to continue moving downward at a lower speed until the contact judgment condition is met. The pressure loading segment drives the locking head to continuously press down at a preset loading rate. The temperature sensor array is arranged at the contact interface between the smart buffer material and the chip, as well as on multiple depth sections along the thickness direction inside the smart buffer material. The pressure sensor array is arranged in a matrix at the upper contact interface between the locking head and the smart buffer material. The displacement sensor is arranged on the drive shaft of the locking head. The temperature sensor array, pressure sensor array, and displacement sensor are synchronously collected with the same sampling period, which is less than one-fifth of the dominant time constant of the locking head mechanical system.

[0079] In step S103, the temperature-mechanical property mapping database is obtained through offline calibration. The offline calibration process performs quasi-static compression test, dynamic mechanical frequency scan test and stress relaxation test on the standard sample of the current batch of smart buffer material to establish the material mechanical parameter mapping relationship with discrete temperature points as index. The current equivalent elastic modulus is used to characterize the stiffness characteristics of the smart buffer material under the current temperature state, and the current equivalent damping coefficient is used to characterize the energy dissipation characteristics of the smart buffer material under the current temperature state.

[0080] In step S104, dynamic compensation correction is used to eliminate the pressure transmission amplitude attenuation and phase delay caused by the viscoelastic properties of the smart buffer material. Displacement data is used to verify the consistency between the pressure field data and the actual motion state of the locking head. The estimated effective pressure distribution on the chip surface is a sequence of compensated pressure values ​​corresponding to the spatial positions of each measuring point in the pressure sensor array.

[0081] In step S105, the global safety risk level is determined by a combination of the uniform compressive stress assessment result and the bending stress assessment result. The uniform compressive stress assessment result is obtained by comparing the area-weighted average effective pressure value with the compressive strength threshold. The bending stress assessment result is calculated based on the thin plate bending theory by combining the spatial gradient of the effective pressure field distribution on the chip surface with the chip thickness and the back support conditions.

[0082] In step S106, the target adjustment displacement in the preventive fine-tuning command is obtained by inverse solution calculation of the multi-physics field coupling calibration model of the locking system, the target adjustment speed is generated by an S-shaped velocity planning curve with continuously changing acceleration, and the response time limit is adaptively set according to the rate of change of the chip's stress state.

[0083] In step S107, after the preventive fine-tuning instruction is executed, the temperature sensor array, pressure sensor array and displacement sensor continuously and synchronously acquire data. After each acquisition, identification, dynamic compensation correction and safe pressure-bearing state assessment are performed in sequence to form a closed-loop adaptive control loop with temperature-pressure dual feedback until the chip's safe pressure-bearing state meets the preset conditions.

[0084] In step S108, the preset conditions include that the area-weighted average effective pressure value reaches the target pressure value and the fluctuation range does not exceed the preset stability tolerance, and that the bending stress values ​​of all monitored micro-elements do not exceed the equivalent stress safety threshold. This embodiment identifies the mechanical state parameters of the intelligent buffer material in real time through temperature field data, and uses these mechanical state parameters to dynamically compensate and correct the pressure field data, thereby obtaining an accurate estimate of the actual pressure-bearing state of the chip surface. Based on this, the safe pressure-bearing state of the chip is assessed, and preventative fine-tuning instructions are generated in the early warning state. This achieves adaptive adjustment of the locking force, avoiding control lag and chip damage risks caused by material temperature dependence.

[0085] This embodiment uses a temperature sensor array to collect temperature field data of the intelligent buffer material in real time. Based on the temperature-mechanical property mapping database, it identifies the current equivalent elastic modulus and the current equivalent damping coefficient online. The identification results are used to dynamically compensate and correct the pressure field data, thereby obtaining an accurate estimate of the actual pressure state of the chip surface. Based on this, the safe pressure state of the chip is evaluated, and preventive fine-tuning instructions are generated in the early warning state. This realizes the adaptive adjustment of the locking force according to the thermodynamic state of the material, avoiding the control lag and chip damage risk caused by the temperature dependence of the material's viscoelasticity, and improving the packaging yield and production efficiency.

[0086] Please see Figure 2 In some embodiments, temperature and pressure field data of the smart buffer material are simultaneously acquired via a temperature sensor array and a pressure sensor array, and displacement data of the locking head is also simultaneously obtained, including:

[0087] S201. Based on the locking process parameters, set the three-stage motion speed curve of the locking head. The three-stage motion speed curve includes a rapid approach stage, a buffer contact stage, and a pressure loading stage.

[0088] S202. During the movement of the locking head according to the three-segment motion speed curve, the temperature signal output by the temperature sensor array, the pressure signal output by the pressure sensor array, and the displacement signal output by the displacement sensor are collected synchronously at a preset sampling period. The preset sampling period is less than one-fifth of the dominant time constant of the mechanical system of the locking head.

[0089] S203. Perform validity checks on the synchronously acquired temperature, pressure and displacement signals respectively. Validity checks include temperature range checks, pressure monotonicity checks and displacement-pressure consistency checks.

[0090] S204. Remove abnormal data that fails the validity check, add a unified timestamp to the remaining valid data after removal, and generate temperature field data, pressure field data, and displacement data.

[0091] In step S201, the rapid approach segment of the three-segment motion speed curve is used to drive the locking head to approach the chip mounting surface at a higher speed. The buffer contact segment is used to continue moving downward at a lower speed after the gap between the end face of the locking head and the upper surface of the smart buffer material reaches a preset approach gap threshold. The pressure loading segment is used to continuously press down at a preset loading rate after the pressure sensor array detects a pressure signal greater than the preset contact judgment threshold.

[0092] In step S202, the temperature signal is the digital temperature value obtained by converting the original potential signal output by each temperature measuring point in the temperature sensor array into a signal conditioning circuit after amplification and filtering; the pressure signal is the digital pressure value obtained by converting the original capacitance change or resistance change output by each pressure measuring point in the pressure sensor array into a signal conditioning circuit; and the displacement signal is the real-time displacement value of the locking head output by the displacement sensor. The preset sampling period is determined based on the dominant time constant of the locking head mechanical system to meet the frequency requirements of the Nyquist sampling theorem for dynamic process sampling.

[0093] In step S203, the temperature range verification is used to determine whether the readings of each temperature measuring point are within the preset reasonable temperature range, the pressure monotonicity verification is used to determine whether the pressure change between adjacent sampling periods meets the monotonically increasing constraint during the pressure loading segment, and the displacement-pressure consistency verification is used to determine whether the pressure value and displacement value at the current moment conform to the force-displacement relationship estimated in the multi-physics coupling calibration model of the locking system.

[0094] In step S204, temperature measurement points that fail the temperature range verification are marked as abnormal measurement points, and the temperature values ​​of abnormal measurement points are replaced by spatial interpolation of adjacent measurement points; pressure signals that fail the pressure monotonicity verification are determined to be electromagnetic interference glitches and are smoothed by median filtering; data acquisition that fails the displacement-pressure consistency verification triggers a sensor fault alarm and suspends the locking process. A unified timestamp is used to ensure the alignment of temperature field data, pressure field data, and displacement data in the time dimension, providing a time reference for subsequent identification of the mechanical state of intelligent buffer materials based on temperature field data and verification of pressure field data based on displacement data.

[0095] This embodiment controls the locking head to complete the approach, contact, and loading process at different speeds using a three-segment motion speed curve. Temperature, pressure, and displacement signals are collected synchronously at a preset sampling period. After validity verification and removal of abnormal data, a unified timestamp is applied, thereby ensuring the alignment and reliability of temperature, pressure, and displacement data in the time dimension. This provides an accurate data foundation for subsequent identification of the mechanical state of intelligent buffer materials based on temperature data and verification of pressure data based on displacement data, avoiding control decision deviations caused by sensor malfunctions or electromagnetic interference.

[0096] Please see Figure 3 In some embodiments, based on temperature field data, a temperature-mechanical property mapping database is queried to identify the current equivalent elastic modulus and equivalent damping coefficient of the smart buffer material, including:

[0097] S301. Spatial interpolation is performed on the discrete temperature values ​​of each temperature measurement point in the temperature field data to reconstruct the continuous two-dimensional temperature field distribution at the contact interface between the intelligent buffer material and the chip.

[0098] S302. Based on the continuous two-dimensional temperature field distribution, the area-weighted average temperature value of the contact interface is calculated using the area of ​​the influence region represented by each temperature measuring point as the weight, and is used as the characteristic temperature at the current moment.

[0099] S303. Using the characteristic temperature as the query index, perform table lookup interpolation in the temperature-mechanical property mapping database to obtain the global equivalent elastic modulus and the global equivalent damping coefficient.

[0100] S304. Calculate the deviation between the temperature value of each temperature measuring point and the characteristic temperature. For measuring points whose deviation exceeds the preset gradient threshold, perform independent lookup table interpolation based on the local temperature of that measuring point to obtain the local mechanical state correction matrix.

[0101] In step S301, spatial interpolation uses either radial basis function interpolation or kriging interpolation. The radial basis function interpolation method selects a Gaussian kernel function or a thin plate spline function as the basis function. The center point of the basis function coincides with the spatial position of each temperature measurement point. The width parameter of the basis function is optimized and determined by cross-validation, so that the reconstructed continuous two-dimensional temperature field distribution strictly passes the measured temperature value at the temperature measurement point position and maintains physical smoothness at non-temperature measurement point positions.

[0102] In step S302, the area of ​​the influence region represented by each temperature measuring point is determined by the Voronoi diagram segmentation method. The area-weighted average temperature value is calculated with the area of ​​the influence region of each temperature measuring point as the weight. The sum of the product of the temperature value of each temperature measuring point and the corresponding area of ​​the influence region is divided by the total area of ​​the contact interface. The result is used as the characteristic temperature at the current moment. The characteristic temperature is used to characterize the overall thermal state of the smart buffer material at the current moment.

[0103] In step S303, the temperature-mechanical property mapping database stores the equivalent elastic modulus, storage modulus, loss modulus, and stress relaxation modulus using discrete temperature points as indexes. When the characteristic temperature is exactly equal to a certain discrete calibration temperature point, the parameter value stored at that temperature point is directly read. When the characteristic temperature falls between two discrete calibration temperature points, the parameter value at that temperature is calculated using the cubic spline interpolation method.

[0104] In step S304, a preset gradient threshold is used to determine the degree of non-uniformity of the temperature field distribution. The area corresponding to the temperature measuring point with a deviation exceeding the preset gradient threshold is identified as a non-uniform temperature field distribution area. The smart buffer material in this area exhibits a different mechanical state from the global state due to temperature differences. The elements at the non-local correction measuring point positions in the local mechanical state correction matrix take the global equivalent elastic modulus and the global equivalent damping coefficient. The local mechanical state correction matrix is ​​used for subsequent local differential dynamic compensation correction of the pressure field data.

[0105] This embodiment reconstructs a continuous two-dimensional temperature field distribution by spatial interpolation of temperature field data, calculates the area-weighted average temperature as the characteristic temperature, and uses the characteristic temperature as the query index to look up and interpolate in the temperature-mechanical property mapping database to obtain the global equivalent elastic modulus and the global equivalent damping coefficient. For measuring points where the temperature deviation exceeds a preset gradient threshold, local temperature is used to independently perform table lookup and interpolation to obtain the local mechanical state correction matrix. This achieves spatial differentiation identification of the mechanical state of the intelligent buffer material, providing mechanical parameters that match the temperature field distribution for the dynamic compensation and correction of subsequent pressure field data, and avoiding local pressure estimation distortion caused by non-uniform temperature field distribution.

[0106] In some embodiments, the deviation between the temperature value of each temperature measuring point and the characteristic temperature is calculated. For measuring points where the deviation exceeds a preset gradient threshold, a lookup table interpolation is performed independently using the local temperature of that measuring point to obtain a local mechanical state correction matrix, including:

[0107] Calculate the temperature deviation between the temperature value of each temperature measuring point and the characteristic temperature, and mark the temperature measuring points whose absolute value of the temperature deviation exceeds the preset gradient threshold as local correction measuring points.

[0108] Using the local temperature of each local correction measurement point as a substitute for the characteristic temperature as a query index, table lookup and interpolation are performed in the temperature-mechanical property mapping database to obtain the local equivalent elastic modulus and local equivalent damping coefficient corresponding to each local correction measurement point.

[0109] Based on the spatial location of each local correction measuring point on the contact interface, the local equivalent elastic modulus and local equivalent damping coefficient are organized into a local mechanical state correction matrix. The elements at the non-local correction measuring point locations in the local mechanical state correction matrix are taken as the global equivalent elastic modulus and global equivalent damping coefficient.

[0110] In this embodiment, the temperature deviation value is calculated based on the characteristic temperature. The difference between the temperature value of each temperature measuring point and the characteristic temperature is calculated one by one. The absolute value of the temperature deviation value is used to measure the degree to which each temperature measuring point deviates from the overall thermal state. The preset gradient threshold is preset according to the sensitivity of the mechanical parameters of the intelligent buffer material to temperature. Materials with higher temperature sensitivity correspond to smaller preset gradient thresholds, and materials with lower temperature sensitivity correspond to larger preset gradient thresholds.

[0111] In this embodiment, the marking of local correction measurement points is used to distinguish regions in the temperature field where the thermal state deviates significantly from the overall temperature. The basis for using local temperature instead of characteristic temperature as the query index is that the smart buffer material in the region where the local correction measurement point is located exhibits different mechanical response characteristics from the global temperature due to temperature differences. If characteristic temperature is still used as the query index, it will mask the difference in mechanical state in this region, resulting in local distortion in subsequent pressure compensation correction.

[0112] In this embodiment, the row and column arrangement of the local mechanical state correction matrix corresponds to the measurement point layout of the pressure sensor array. The elements at each position in the matrix correspond one-to-one with the chip surface area covered by the pressure measurement point at that position. The elements at the local correction measurement point positions are the local equivalent elastic modulus and local equivalent damping coefficient obtained by independent lookup table interpolation, while the elements at the non-local correction measurement point positions are the global equivalent elastic modulus and global equivalent damping coefficient. The local mechanical state correction matrix is ​​used to apply compensation parameters that match the mechanical state at different spatial positions to the pressure measurement points in the subsequent dynamic compensation and correction process.

[0113] This embodiment calculates the deviation between each temperature measuring point and the characteristic temperature and marks the local correction measuring points. The local equivalent elastic modulus and local equivalent damping coefficient are obtained by interpolating the table using the local temperature instead of the characteristic temperature. These are then organized into a local mechanical state correction matrix according to their spatial location. This enables independent identification of the mechanical state in areas with non-uniform temperature field distribution. It provides a basis for applying differentiated compensation parameters to the pressure measuring points corresponding to the spatial location of the local correction measuring points, thus avoiding the problem of local pressure compensation mismatch caused by temperature gradient.

[0114] In some embodiments, based on the current equivalent elastic modulus, equivalent damping coefficient, and displacement data, dynamic compensation and correction are performed on the pressure field data to generate an estimate of the effective pressure distribution on the chip surface, including:

[0115] The force transmission process of the smart buffer material is equivalent to a first-order inertial element, and the time constant of the first-order inertial element is determined by the ratio of the current equivalent damping coefficient to the current equivalent elastic modulus.

[0116] For the pressure value sequence of each pressure measuring point in the pressure field data, dynamic compensation calculation is performed using the inverse model of the first-order inertial element. The dynamic compensation calculation is performed in the discrete time domain. The compensated pressure value at the current moment is determined by the original pressure value at the current moment, the compensated pressure value at the previous moment, and the current equivalent elastic modulus and the current equivalent damping coefficient.

[0117] For the pressure measurement points corresponding to the spatial locations of each local correction measurement point in the local mechanical state correction matrix, the local equivalent elastic modulus and local equivalent damping coefficient corresponding to the local correction measurement point are used to replace the current equivalent elastic modulus and current equivalent damping coefficient, and dynamic compensation calculation is performed independently.

[0118] The compensated pressure values ​​of each pressure measuring point are arranged according to their spatial location to generate an estimated value of the effective pressure distribution on the chip surface.

[0119] In this embodiment, the force transmission process of the smart buffer material is represented by a transfer function in the frequency domain. The gain of the transfer function is equal to the ratio of the current equivalent elastic modulus to the initial elastic modulus of the smart buffer material. The phase lag of the transfer function is the damping angle, which is determined by the ratio of the current equivalent damping coefficient to the current equivalent elastic modulus. The first-order inertial element is the equivalent form of the transfer function in the time domain. The time constant of the first-order inertial element is updated in real time with the changes in the current equivalent elastic modulus and the current equivalent damping coefficient.

[0120] In this embodiment, the dynamic compensation operation is based on the inverse model of the first-order inertial element. The inverse model operation is implemented in the discrete time domain using a first-order difference equation. The compensated pressure value at the current moment is obtained by a weighted combination of the original pressure value at the current moment and the compensated pressure value at the previous moment. The weight of the weighted combination is determined by the current equivalent elastic modulus and the current equivalent damping coefficient. When the current equivalent elastic modulus increases, the weight of the original pressure value at the current moment increases. When the current equivalent damping coefficient increases, the weight of the compensated pressure value at the previous moment increases.

[0121] In this embodiment, the pressure measurement points corresponding to the spatial locations of the local correction measurement points in the local mechanical state correction matrix have different mechanical states than the global identification results because the temperature of the smart buffer material at that location deviates significantly from the characteristic temperature. If the dynamic compensation calculation is still performed using the global parameters, the pressure compensation accuracy at that location will decrease due to the mismatch of the mechanical state parameters. Therefore, the local equivalent elastic modulus and the local equivalent damping coefficient are used to replace the global parameters to perform the dynamic compensation calculation independently, so that the pressure compensation result at that location matches the local mechanical state.

[0122] In this embodiment, the compensated pressure value corresponding to each pressure measuring point in the estimated effective pressure distribution on the chip surface is used to characterize the actual pressure-bearing state at the interface between the lower surface of the smart buffer material and the chip. This estimated value serves as a direct input for subsequent evaluation of the chip's safe pressure-bearing state. The spatial arrangement of the compensated pressure values ​​at each pressure measuring point is consistent with the measuring point layout of the pressure sensor array to ensure that the pressure-bearing state at each location on the chip surface can be accurately corresponded.

[0123] This embodiment equates the force transmission process of the smart buffer material to a first-order inertial element, determines the time constant using the current equivalent elastic modulus and the current equivalent damping coefficient, performs inverse model dynamic compensation calculations on the pressure field data point by point, and independently performs compensation using local mechanical parameters for locally corrected measurement points. This eliminates the pressure transmission amplitude attenuation and phase delay caused by the viscoelastic properties of the smart buffer material, and obtains an effective pressure distribution estimate that matches the actual pressure-bearing state of the chip surface, providing accurate input data for subsequent chip safety pressure-bearing state assessment.

[0124] In some embodiments, the pressure value sequence of each pressure measuring point in the pressure field data is dynamically compensated using the inverse model of a first-order inertial element, including:

[0125] The inverse model of the first-order inertial element is discretized in the discrete time domain to obtain the compensation operation rules in the form of first-order difference equations.

[0126] Based on the current equivalent elastic modulus and the current equivalent damping coefficient, the compensation coefficient of the first-order difference equation is updated in real time. The compensation coefficient is inversely proportional to the time constant of the first-order inertial element.

[0127] For each pressure measuring point, the compensation calculation rules are executed on a sampling cycle basis:

[0128] The compensated pressure value of the current sampling period is equal to the weighted combination of the original pressure value of the current sampling period and the compensated pressure value of the previous sampling period. The weight of the weighted combination is determined by the compensation coefficient updated in real time.

[0129] When the current equivalent elastic modulus increases, the compensation coefficient increases accordingly, which increases the weight of the original pressure value in the current sampling period in the weighted combination.

[0130] When the current equivalent damping coefficient increases, the compensation coefficient decreases accordingly, thereby increasing the weight of the compensated pressure value from the previous sampling period in the weighted combination.

[0131] In this embodiment, the discretization process of the inverse model of the first-order inertial element in the discrete time domain determines the specific form of the difference equation based on the relative relationship between the sampling period and the time constant of the first-order inertial element. The compensation operation rules of the difference equation form are used to describe the recursive relationship between the compensated pressure value of the current sampling period, the original pressure value of the current sampling period, and the compensated pressure value of the previous sampling period.

[0132] In this embodiment, the real-time update of the compensation coefficient is performed based on the identification results of the current equivalent elastic modulus and the current equivalent damping coefficient. The current equivalent elastic modulus and the current equivalent damping coefficient change dynamically with the change of temperature field data. The compensation coefficient increases as the time constant of the first-order inertial element decreases and decreases as the time constant of the first-order inertial element increases, thereby ensuring that the response speed of the compensation calculation matches the real-time mechanical state of the intelligent buffer material.

[0133] In this embodiment, when the compensation operation rule is executed periodically, the compensated pressure value of each sampling period is used as the input for the compensation operation of the next sampling period, forming a recursive operation chain. The compensated pressure value at the initial moment of the recursive operation chain is taken as the original pressure value at that moment. The operation process of the recursive operation chain continues until all sampling periods of the pressure value sequence have been processed.

[0134] In this embodiment, the sum of the weights of the original pressure value in the current sampling period and the compensated pressure value in the previous sampling period in the weighted combination is a fixed value. When the current equivalent elastic modulus increases, the stiffness of the intelligent buffer material increases, the pressure transmission attenuation decreases, and the contribution of the original pressure value in the current sampling period to the compensation result increases. When the current equivalent damping coefficient increases, the viscosity of the intelligent buffer material increases, the pressure transmission phase delay increases, and the smoothing effect of the compensated pressure value in the previous sampling period on the compensation result is enhanced, thereby compensating for amplitude attenuation while also correcting phase delay.

[0135] This embodiment discretizes the inverse model of the first-order inertial element into a compensation operation rule in the form of a first-order difference equation. The compensation coefficient is updated in real time according to the current equivalent elastic modulus and the current equivalent damping coefficient. The recursive compensation operation is performed on a sampling cycle basis, thereby realizing the adaptive adjustment of the compensation parameters with the real-time mechanical state changes of the intelligent buffer material. This avoids the compensation deviation caused by the fixed compensation coefficient when the material state changes, and improves the pressure compensation accuracy and response speed.

[0136] In some embodiments, based on the estimated effective pressure distribution on the chip surface and combined with the equivalent stress safety threshold, the chip's safe pressure-bearing state is assessed, and a global safety risk level identifier is generated, including:

[0137] The estimated effective pressure distribution on the chip surface is mapped to each monitoring micro-element on the chip surface. Spatial interpolation is performed on the monitoring micro-element that is not directly covered by the pressure measurement point to generate a continuous effective pressure field distribution covering the entire chip surface.

[0138] Calculate the area-weighted average effective pressure value of the continuous effective pressure field distribution, compare the area-weighted average effective pressure value with the compressive strength threshold, and generate a uniform compressive stress assessment result;

[0139] Based on the spatial gradient of the continuous effective pressure field distribution, combined with the chip thickness and back support conditions, the bending stress of each monitoring micro-element is calculated according to the thin plate bending theory. The bending stress is compared with the equivalent stress safety threshold to generate the bending stress assessment result.

[0140] Based on the results of the uniform compressive stress assessment and the bending stress assessment, the overall safety risk level is determined. The overall safety risk level includes three levels: safe state, warning state, and dangerous state.

[0141] In this embodiment, the pressure values ​​of each pressure measuring point in the estimated effective pressure distribution on the chip surface are compensated and mapped to the corresponding monitoring micro-element based on the spatial projection position of each pressure measuring point on the contact interface. The monitoring micro-element not directly covered by the pressure measuring point is estimated based on the pressure value of the adjacent assigned monitoring micro-element through bilinear interpolation or bicubic interpolation. The continuous effective pressure field distribution is used to characterize the true pressure state at each location on the chip surface.

[0142] In this embodiment, the area-weighted average effective pressure value is calculated with the area represented by each monitoring micro-element as the weight. The effective pressure value of each monitoring micro-element is multiplied by the corresponding area and then divided by the total area of ​​the chip surface. The compressive strength threshold decreases as the temperature increases based on the high-temperature softening characteristics of silicon material. The uniform compressive stress assessment result is used to determine whether the chip has an overall crushing risk.

[0143] In this embodiment, the spatial gradient is calculated using the central difference method. The first pressure gradient is calculated by taking the effective pressure values ​​of three adjacent monitoring micro-elements along the first horizontal direction, and the second pressure gradient is calculated by taking the effective pressure values ​​of three adjacent monitoring micro-elements along the second horizontal direction. The first horizontal direction and the second horizontal direction are orthogonal. The bending stress is determined by the first pressure gradient, the second pressure gradient, the chip thickness, and the back support conditions based on the thin plate bending theory. The bending stress assessment result is used to determine whether there is a risk of local fracture in the chip.

[0144] In this embodiment, the global safety risk level indicator is determined based on the ratio of the area-weighted average effective pressure value to the compressive strength threshold and the ratio of the maximum bending stress value to the equivalent stress safety threshold. When neither ratio exceeds the corresponding threshold, it is determined to be in a safe state. When either ratio exceeds the corresponding threshold but does not exceed the preset over-limit tolerance, it is determined to be in a warning state. When either ratio exceeds the preset over-limit tolerance, it is determined to be in a dangerous state. The global safety risk level indicator is used to trigger the corresponding locking force adjustment strategy.

[0145] This embodiment maps the estimated effective pressure distribution on the chip surface to each monitoring micro-element and interpolates to generate a continuous effective pressure field distribution. It calculates the comparison results of the area-weighted average effective pressure value and the compressive strength threshold, as well as the comparison results of the bending stress and the equivalent stress safety threshold based on spatial gradient and thin plate bending theory. It comprehensively determines the global safety risk level, thereby realizing a comprehensive assessment of the chip's safe pressure-bearing state from two dimensions: uniform compressive stress and bending stress. This avoids ignoring the risk of local fracture caused by uneven pressure distribution by using only a single pressure threshold as a protection criterion.

[0146] In some embodiments, based on the spatial gradient of the continuous effective pressure field distribution, combined with the chip thickness and back support conditions, the bending stress of each monitoring micro-element is calculated according to the thin-plate bending theory, including:

[0147] For each monitoring micro-element, the effective pressure values ​​of the two monitoring micro-elements adjacent to the monitoring micro-element are taken along the first horizontal direction, and the first pressure gradient of the monitoring micro-element in the first horizontal direction is calculated by the central difference method.

[0148] The effective pressure values ​​of two monitoring micro-elements adjacent to the monitoring micro-element are taken along the second horizontal direction. The second pressure gradient of the monitoring micro-element in the second horizontal direction is calculated by the central difference method. The second horizontal direction is orthogonal to the first horizontal direction.

[0149] Based on the first pressure gradient and the second pressure gradient, combined with the chip thickness and the back support conditions, the first bending stress component and the second bending stress component in the second horizontal direction of the monitoring micro-element are calculated according to the thin plate bending theory.

[0150] The bending stress value of the monitoring micro-element is obtained by vector synthesis of the first bending stress component and the second bending stress component.

[0151] In this embodiment, the first horizontal direction and the second horizontal direction constitute an orthogonal coordinate system on the chip surface. The center difference method takes the monitoring micro-element as the center, takes the ratio of the difference of the effective pressure values ​​of two monitoring micro-elements adjacent to the monitoring micro-element in the first horizontal direction to the distance between the two monitoring micro-elements as the first pressure gradient, and takes the ratio of the difference of the effective pressure values ​​of two monitoring micro-elements adjacent to the monitoring micro-element in the second horizontal direction to the distance between the two monitoring micro-elements as the second pressure gradient.

[0152] In this embodiment, the back support condition is used to characterize the contact constraint state between the back of the chip and the supporting structure. When the back support condition is different, the degree of bending deformation of the chip under the same pressure gradient is different. The chip thickness is used to characterize the bending stiffness of the chip cross section. The larger the chip thickness, the smaller the bending stress generated under the same pressure gradient. The calculation of the first bending stress component and the second bending stress component are both based on the thin plate bending theory, which converts the pressure gradient into bending stress inside the chip.

[0153] In this embodiment, the vector synthesis uses the first bending stress component and the second bending stress component as orthogonal components. The synthesized bending stress value is used to characterize the comprehensive bending stress state of the monitoring micro-element under the combined action of bidirectional pressure gradient. The synthesized bending stress value is compared with the equivalent stress safety threshold to determine whether the monitoring micro-element has a risk of fracture.

[0154] This embodiment calculates the pressure gradient along the first and second horizontal directions using the center difference method. Combining the chip thickness and back support conditions, it calculates the bending stress components in the two orthogonal directions based on the thin plate bending theory and then vector synthesizes them. This yields the comprehensive bending stress value of each monitoring micro-element under the combined action of the bidirectional pressure gradient, providing an accurate stress basis for determining whether the chip has a risk of local fracture. This avoids the problem of underestimating the actual stress state of the chip by only considering bending stress in one direction.

[0155] In some embodiments, when the global security risk level is identified as a warning state, a preventative fine-tuning instruction is generated. This preventative fine-tuning instruction includes the target adjustment displacement and target adjustment speed of the locking head, and includes:

[0156] The difference between the estimated effective pressure distribution on the chip surface and the target pressure distribution is calculated and used as the desired pressure adjustment amount.

[0157] Based on the inverse solution calculation of the multiphysics coupling calibration model of the locking system, combined with the current equivalent elastic modulus, the target adjustment displacement of the locking head is estimated according to the expected pressure adjustment amount.

[0158] Based on the target adjustment displacement, an S-shaped velocity planning curve with continuously varying acceleration is used to generate the target adjustment speed of the locking head;

[0159] Based on the rate of change of the current chip stress state, the response time limit is adaptively set to generate a preventive fine-tuning instruction that includes the target adjustment displacement, the target adjustment speed, and the response time limit.

[0160] In this embodiment, the target pressure distribution is the distribution of the target pressure value in the latching process parameters on each monitoring micro-element on the chip surface. The target pressure value at each monitoring micro-element in the target pressure distribution is preset according to the chip specifications and process requirements. The difference between the estimated effective pressure distribution on the chip surface and the target pressure distribution is calculated point by point at the corresponding monitoring micro-element position, and the resulting difference sequence is used as the expected pressure adjustment amount.

[0161] In this embodiment, the inverse operation of the multiphysics coupling calibration model of the locking system takes the desired pressure adjustment amount as input and the displacement adjustment amount of the locking head as output. During the inverse operation, the current equivalent elastic modulus is used to establish the mechanical relationship between the desired pressure adjustment amount and the target adjustment displacement amount. The estimation of the target adjustment displacement amount is performed based on the linear constitutive relationship between force and displacement. The larger the current equivalent elastic modulus, the smaller the target adjustment displacement amount required to generate the same desired pressure adjustment amount.

[0162] In this embodiment, the acceleration of the S-shaped speed planning curve gradually increases from zero to a peak value in the initial stage, remains constant in the middle stage, and gradually decreases to zero in the final stage. The continuous change in acceleration is used to avoid inertial impact caused by sudden acceleration changes during the adjustment process of the locking head. The target adjustment speed is determined by the target adjustment displacement and the response time limit. The larger the target adjustment displacement or the shorter the response time limit, the larger the peak value of the target adjustment speed.

[0163] In this embodiment, the rate of change of the current chip stress state is determined by comparing the change in the maximum bending stress value in two adjacent safe pressure state assessment results. When the rate of change is large, the response time is shortened to ensure timely adjustment, and when the rate of change is small, the response time is extended to ensure smooth adjustment. After the preventive fine-tuning instruction is generated, it is parsed by the control system into a control signal sequence for the locking head drive motor and executed.

[0164] This embodiment calculates the difference between the estimated effective pressure distribution on the chip surface and the target pressure distribution to obtain the desired pressure adjustment amount. Based on the inverse solution operation of the multi-physics coupling calibration model of the locking system and combined with the current equivalent elastic modulus, the target adjustment displacement is estimated. The target adjustment speed is generated by using an S-shaped velocity planning curve, and the response time limit is adaptively set according to the rate of change of the chip's stress state. This generates a preventive fine-tuning command that matches the real-time mechanical state of the intelligent buffer material, realizing proactive predictive adjustment in the early warning state and avoiding overshoot or undershoot problems caused by fixed gain adjustment.

[0165] In some embodiments, the movement speed curve of the locking head is set according to the locking process parameters, and the locking process is started, including:

[0166] Based on the specification identification code of the chip to be mounted, retrieve the corresponding latching process parameters from the process database;

[0167] During the rapid approach phase, the locking head is controlled to move toward the chip mounting surface at a first speed until the gap between the end face of the locking head and the upper surface of the smart buffer material is reduced to a preset approach gap threshold.

[0168] When the gap decreases to a preset gap threshold, switch to the buffer contact section and control the locking head to continue descending at a second speed lower than the first speed until the pressure sensor array detects a pressure signal greater than the preset contact determination threshold.

[0169] When the pressure sensor array detects a pressure signal greater than the preset contact determination threshold, it switches to the pressure loading stage and controls the locking head to continue pressing down at the preset loading rate until the target pressure value or target displacement value is reached, and the locking process is started.

[0170] In this embodiment, the specification identification code is the unique identification code of the chip to be mounted, and the process database is a storage space that stores the mounting process parameters corresponding to chips of different specifications. The mounting process parameters include target pressure value, target displacement value, speed curve parameters, and equivalent stress safety threshold of the monitoring micro-element on the chip surface. The correspondence between the specification identification code and the mounting process parameters is established in advance before the mounting equipment is put into use.

[0171] In this embodiment, the first speed is the high-speed movement speed of the locking head in the rapid approach phase, which is used to shorten the time it takes for the locking head to move from the initial waiting position to the vicinity of the chip mounting surface. The preset approach gap threshold is the minimum safe distance allowed between the end face of the locking head and the upper surface of the smart buffer material. When the gap between the end face of the locking head and the upper surface of the smart buffer material decreases to the preset approach gap threshold, the locking head is about to come into contact with the smart buffer material. At this time, the movement speed needs to be reduced to avoid impact.

[0172] In this embodiment, the second speed is the low-speed movement speed of the locking head in the buffer contact section. The second speed is lower than the first speed and is used to ensure the stability of the locking head when it comes into contact with the smart buffer material. The preset contact judgment threshold is the pressure value at which the pressure sensor array determines that the locking head and the smart buffer material have established contact. When the pressure sensor array detects a pressure signal greater than the preset contact judgment threshold, it indicates that the end face of the locking head has formed effective contact with the upper surface of the smart buffer material, and at this time, the pressure loading stage can be entered.

[0173] In this embodiment, the preset loading rate is the displacement change rate of the locking head in the pressure loading section. The target pressure value and the target displacement value are used as the termination conditions of the pressure loading section. The locking head continues to press down at the preset loading rate. When the pressure value detected by the pressure sensor array reaches the target pressure value or the displacement value detected by the displacement sensor reaches the target displacement value, the locking head stops pressing down, the locking process is officially started, and then the synchronous acquisition stage of temperature field data and pressure field data is entered.

[0174] In this embodiment, the corresponding locking process parameters are retrieved from the process database using the specification identification code. The locking head is controlled to sequentially pass through the rapid approach section, the buffer contact section, and the pressure loading section. The transition between each section is based on the gap threshold and the pressure signal, thereby achieving a smooth transition in the contact process between the locking head and the intelligent buffer material. This avoids impact damage caused by high-speed contact and provides stable mechanical motion conditions for the subsequent synchronous acquisition of temperature field data and pressure field data.

[0175] In some embodiments, before obtaining the locking process parameters, the method further includes:

[0176] A multiphysics coupling calibration model for the locking system is established. This model is used to describe the thermal-mechanical coupling transmission relationship between the locking head, the smart buffer material, and the bare die chip.

[0177] The geometric configuration of the locking head, intelligent buffer material and bare die chip stacked vertically is abstracted into a three-layer series structure. In the three-layer series structure, the locking head is the active driving layer, the intelligent buffer material is the force transmission and energy dissipation layer, and the bare die chip is the passive pressure bearing layer.

[0178] A rigid body motion equation is established for the locking head, and the transmission relationship between the output displacement of the locking head and the control voltage of the drive motor is expressed as a series connection of a first-order inertial element and an integral element.

[0179] A temperature-dependent viscoelastic constitutive equation is established for smart buffer materials. A modified Maxwell-Kelvin series model is used to describe the mechanical behavior of smart buffer materials under the combined action of compressive and thermal loads. In the modified Maxwell-Kelvin series model, the elastic coefficient and viscosity coefficient of each element are set as functions of temperature.

[0180] A brittle fracture mechanics criterion model was established for bare die chips. The chip surface was divided into multiple monitoring micro-elements along the horizontal direction. Based on the Weber brittleness statistical theory of silicon materials, the mapping relationship between the equivalent stress of each monitoring micro-element and the overall failure probability of the chip was established. The equivalent stress safety threshold of each monitoring micro-element was determined based on the maximum failure probability allowed by the packaging process.

[0181] The equations of motion of the locking head rigid body, the temperature-dependent viscoelastic constitutive equation of the smart buffer material, and the brittle fracture mechanics criterion equation of the bare die chip are coupled together to obtain a multi-physics coupling calibration model of the locking system. The calibration model takes the driving control voltage of the locking head and the temperature field distribution of the smart buffer material as input variables, and the equivalent stress distribution of each monitoring micro-element on the surface of the bare die chip as output variables.

[0182] In this embodiment, the abstract of the three-layer series structure is based on the force transmission path in the vertical direction of the locking head, the intelligent buffer material, and the bare die chip. The active driving layer is used to convert the control voltage of the drive motor into the mechanical displacement output of the locking head. The force transmission and energy dissipation layer is used to transmit pressure and dissipate impact energy between the locking head and the bare die chip. The passive pressure-bearing layer is used to withstand the effective pressure transmitted by the intelligent buffer material.

[0183] In this embodiment, the time constant of the first-order inertial element in the rigid body motion equation is jointly determined by the electromagnetic time constant of the drive motor and the equivalent inertia of the mechanical transmission chain. The gain of the integral element is determined by the lead screw and the reduction ratio. The rigid body motion equation is used to describe the dynamic response relationship between the control voltage of the drive motor and the output displacement of the locking head.

[0184] In this embodiment, the modified Maxwell-Kelvin cascade model consists of a set of Maxwell units and a set of Kelvin units connected in series. The Maxwell units are used to characterize the instantaneous elastic response and stress relaxation behavior of the smart buffer material, while the Kelvin units are used to characterize the delayed elastic response and creep behavior of the smart buffer material. The functional relationship between the elastic coefficient and viscosity coefficient of each unit and temperature is obtained through offline experimental calibration.

[0185] In this embodiment, the division of monitoring micro-elements is determined based on the geometric dimensions of the chip, and the equivalent stress safety threshold of each monitoring micro-element is determined based on the mapping relationship between the overall failure probability of the chip and the Weberian brittleness statistical theory. The failure probability increases with the increase of equivalent stress, and the failure probability corresponding to the equivalent stress safety threshold is the maximum failure probability allowed by the packaging process.

[0186] In this embodiment, the multi-physics coupling calibration model of the coupled locking system is used to describe the entire link transmission relationship from the driving control voltage and temperature field distribution to the equivalent stress distribution of each monitoring micro-element on the surface of the bare die chip. During the locking process, the calibration model is used to update the force transmission parameters based on the currently identified equivalent elastic modulus and equivalent damping coefficient, and in the early warning state, it estimates the target adjustment displacement based on the expected pressure adjustment amount through inverse solution calculation.

[0187] This embodiment abstracts the locking head, intelligent buffer material, and bare die chip into a three-layer series structure, and establishes rigid body motion equations, temperature-dependent viscoelastic constitutive equations, and brittle fracture mechanics criterion equations in parallel and coupled. This obtains the full-link transmission relationship from the driving control voltage and temperature field distribution to the equivalent stress distribution on the chip surface, providing a model basis for updating pressure transmission parameters during the locking process and inverse calculation of the target adjustment displacement under early warning conditions.

[0188] In some embodiments, executing the preventative fine-tuning instructions further includes:

[0189] The preventive fine-tuning command is parsed into a control signal sequence for the locking head drive motor. The generation of the control signal sequence adopts a model reference adaptive control framework.

[0190] In the model reference adaptive control framework, a reference model is set. The reference model describes the desired dynamic relationship between the displacement command of the locking head and the effective pressure response of the chip surface under ideal conditions. The desired dynamic relationship is designed as a second-order system with a preset damping ratio and a preset natural frequency.

[0191] Based on the currently identified mechanical state parameters of the intelligent buffer material, the control gain of the adjustable controller is adjusted in real time. The adjustment rule for the control gain is as follows: when the identified equivalent elastic modulus increases, the proportional control gain is reduced accordingly; when the identified equivalent damping coefficient decreases, the derivative control gain is increased accordingly.

[0192] During the execution of the adjustment command, real-time acquisition of temperature field data and pressure field data is continuously performed, and the identification, compensation correction and evaluation steps are repeatedly executed to form a closed-loop adaptive control loop with temperature-pressure dual feedback.

[0193] After each complete operation cycle from data acquisition to adjustment command execution, the adjustment effect is evaluated. The adjustment effect evaluation is achieved by comparing the changing trends of the maximum equivalent stress value and the area-weighted average effective pressure value on the chip surface before and after adjustment.

[0194] If the maximum equivalent stress value after adjustment does not show a downward trend or the safety risk level does not decrease within the preset response time limit, the adjustment strategy recalculation is triggered. The adjustment strategy recalculation includes re-evaluating the mechanical state identification results of the smart buffer material, and increasing the adjustment range or switching to a more conservative adjustment strategy if the identification results are confirmed to be reliable.

[0195] In this embodiment, the reference model in the model reference adaptive control framework is used to provide the desired dynamic response benchmark for the adjustable controller. The values ​​of the preset damping ratio and the preset natural frequency are determined based on the inherent characteristics of the locking head mechanical system. The preset natural frequency is not higher than the natural frequency of the locking head mechanical system to avoid exciting structural resonance. The deviation between the output of the reference model and the effective pressure response on the chip surface serves as the basis for the adjustable controller to adjust the control gain.

[0196] In this embodiment, the control gain adjustment of the adjustable controller is performed based on the currently identified equivalent elastic modulus and equivalent damping coefficient. When the equivalent elastic modulus increases, the stiffness of the intelligent buffer material increases, and the open-loop gain of the system increases accordingly. At this time, the proportional control gain is reduced to avoid oscillation caused by the increase in system stiffness. When the equivalent damping coefficient decreases, the energy dissipation capacity of the intelligent buffer material weakens, and the system damping decreases accordingly. At this time, the differential control gain is increased to provide additional damping compensation.

[0197] In this embodiment, each complete operation cycle in the closed-loop adaptive control loop includes the synchronous acquisition of temperature field data and pressure field data, online identification of equivalent elastic modulus and equivalent damping coefficient, dynamic compensation and correction of pressure field data, evaluation of chip safe pressure bearing state, and execution of adjustment instructions. The temperature field data and pressure field data after the adjustment instructions are executed serve as the input for the next operation cycle, forming a continuous feedback adjustment process.

[0198] In this embodiment, the downward trend of the maximum equivalent stress value in the adjustment effect evaluation is determined by comparing the relative change of the maximum equivalent stress value in two adjacent evaluation results. The trend of the area-weighted average effective pressure value is used to determine whether the adjustment direction is correct. If the maximum equivalent stress value shows a downward trend after adjustment and the global safety risk level indicator changes from a high level to a low level, then the current adjustment direction is determined to be correct, and the remaining adjustment instructions are executed until the target state is reached.

[0199] In this embodiment, the triggering condition for recalculating the adjustment strategy is that the maximum equivalent stress value after adjustment does not show a downward trend or the safety risk level does not decrease within the preset response time limit. During the recalculation of the adjustment strategy, the temperature sensor array and pressure sensor array are first checked for abnormalities to rule out the possibility of sensor failure causing the identification result to be distorted. On the premise of confirming that the identification result is reliable, the adjustment range is increased or a more conservative adjustment strategy is switched. The more conservative adjustment strategy includes extending the adjustment response time, reducing the single adjustment displacement, and triggering the emergency protection mode in advance.

[0200] This embodiment uses a model reference adaptive control framework to parse preventive fine-tuning commands into a sequence of control signals. The control gain is adjusted in real time based on the identified mechanical state parameters. During the adjustment process, the acquisition, identification, compensation correction, and evaluation steps are continuously executed to form a closed-loop control loop. The adjustment effect is evaluated to determine whether the adjustment direction is correct and triggers the recalculation of the adjustment strategy when necessary. This achieves continuous optimization and adaptive adjustment of the locking force adjustment process, avoiding the adjustment failure problem caused by fixed control parameters when the material state changes.

[0201] In some embodiments, based on the estimated effective pressure distribution on the chip surface and combined with the equivalent stress safety threshold, the safe pressure-bearing state of the chip is assessed, and a global safety risk level identifier is generated, which further includes:

[0202] When the global safety risk level is identified as Level 1 safety status, the current movement state of the locking head remains unchanged, and the locking process continues to be executed according to the preset speed curve and pressure target value, while maintaining real-time monitoring of the temperature-pressure dual feedback signals;

[0203] When the global safety risk level is identified as Level 3 dangerous state, the emergency protection mode is activated. The emergency protection mode first immediately stops the downward movement of the locking head, and then determines whether to perform the locking head retraction operation and the retraction distance based on the quantification value of the degree of exceeding the limit and the coordinate set of local high-risk areas.

[0204] The calculation basis for the back-off distance is: the amount of material springback required to reduce the current maximum equivalent stress value to below the equivalent stress safety threshold, plus the reserved safety margin back-off amount;

[0205] In emergency protection mode, an alarm signal is triggered simultaneously to report the current abnormal status to the upper control system, and all temperature-pressure dual-channel sensor data within a preset time window before and after the time of the abnormality are recorded for subsequent fault analysis.

[0206] In this embodiment, under the first-level safety condition, the locking head maintains its current motion state. The preset speed curve and pressure target value serve as the benchmark for the continued execution of the locking process. Real-time monitoring of temperature-pressure dual feedback signals is used to capture the dynamic changes that may occur in the chip's safe pressure-bearing state. During the monitoring process, the acquisition frequency of temperature field data and pressure field data is consistent with the acquisition frequency during the normal execution of the locking process.

[0207] In this embodiment, the downward movement of the locking head is immediately stopped under the third-level danger state. The over-limit quantification value is used to characterize the degree to which the current maximum equivalent stress value exceeds the equivalent stress safety threshold. The local high-risk area coordinate set is used to locate the monitoring micro-element where the equivalent stress value on the chip surface exceeds the equivalent stress safety threshold. Whether the back-off operation is executed and the determination of the back-off distance are based on the over-limit quantification value and the local high-risk area coordinate set.

[0208] In this embodiment, the material springback amount is determined based on the current equivalent elastic modulus and the quantification value of the degree of exceeding the limit. The larger the current equivalent elastic modulus, the smaller the material springback amount required to reduce the current maximum equivalent stress value to below the equivalent stress safety threshold. The safety margin backoff amount is used to provide additional safety margin for the stress state after the locking head retracts. The value of the safety margin backoff amount is preset based on chip specifications and process experience.

[0209] In this embodiment, the alarm signal is triggered to notify the upper control system that the current locking process is in an abnormal state. All temperature-pressure dual-channel sensing data within the preset time window includes the temperature value of each measuring point of the temperature sensor array, the pressure value of each measuring point of the pressure sensor array, and the displacement value output by the displacement sensor. All temperature-pressure dual-channel sensing data within the preset time window before and after the abnormality occurs are recorded and stored for subsequent fault analysis to trace the cause and evolution of the abnormality.

[0210] This embodiment maintains the movement of the locking head and keeps it under real-time monitoring in the first-level safety state. In the third-level danger state, it immediately stops the downward movement and determines whether to retreat and the retreat distance based on the quantified value of the degree of exceeding the limit and the coordinate set of the local high-risk area. At the same time, it triggers an alarm signal and records all sensor data within a preset time window before and after the abnormality. This realizes graded response protection for the chip's safe pressure state, avoids irreversible damage to the chip caused by continued downward pressure in the danger state, and provides complete data support for subsequent fault analysis.

[0211] The aforementioned technical solution acquires temperature field data in real time through a temperature sensor array. Based on a temperature-mechanical property mapping database, it identifies the current equivalent elastic modulus and equivalent damping coefficient of the intelligent buffer material online and uses this data for dynamic compensation and correction of the pressure field data. This generates an estimate of the effective pressure distribution on the chip surface, solving the control lag problem caused by the disconnect between temperature control and force control in existing technologies and the inability to perceive real-time changes in the material's mechanical state. Simultaneously, based on the compensated effective pressure distribution, the chip's safe pressure-bearing state is assessed. In the early warning state, preventative fine-tuning commands are generated, realizing a shift from passive emergency braking to proactive predictive protection. This eliminates control lag and overshoot caused by the thermo-mechanical coupling effect, significantly improving the response speed and adjustment accuracy of the latching force control, reducing the probability of chip microcracks and edge breakage, and improving packaging yield.

[0212] Finally, it should be noted that although the above embodiments have been described in the text and drawings of this application, this should not limit the scope of patent protection of this application. Any technical solutions that are based on the essential concept of this application and utilize the content described in the text and drawings of this application, resulting in equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this application.

Claims

1. A locking force adaptive control method based on temperature-pressure dual feedback, characterized in that, include: Obtain the latching process parameters, which include the target pressure value, the target displacement value, and the equivalent stress safety threshold of the monitoring micro-element on the chip surface; The motion speed curve of the locking head is set according to the locking process parameters, the locking process is started, and the temperature field data and pressure field data of the intelligent buffer material are collected synchronously through the temperature sensor array and pressure sensor array, and the displacement data of the locking head is obtained synchronously. Based on the temperature field data, the temperature-mechanical property mapping database is queried to identify the current equivalent elastic modulus and equivalent damping coefficient of the smart buffer material; Based on the current equivalent elastic modulus, equivalent damping coefficient, and displacement data, the pressure field data is dynamically compensated and corrected to generate an estimated value of the effective pressure distribution on the chip surface. Based on the estimated effective pressure distribution on the chip surface and the equivalent stress safety threshold, the chip's safe pressure-bearing state is assessed, and a global safety risk level identifier is generated. When the global security risk level is identified as a warning state, a preventive fine-tuning instruction is generated, which includes the target adjustment displacement and target adjustment speed of the locking head. The preventive fine-tuning command is executed, and the acquisition, identification, compensation correction and evaluation steps are repeated until the chip's safe pressure-bearing state meets the preset conditions; When the chip's safe pressure-bearing state meets the preset conditions, the locking is determined to be complete and the locking result is output.

2. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 1, characterized in that, Temperature and pressure field data of the intelligent buffer material are simultaneously acquired using temperature and pressure sensor arrays, and displacement data of the locking head is also simultaneously obtained, including: Based on the locking process parameters, a three-segment movement speed curve for the locking head is set, which includes a rapid approach segment, a buffer contact segment, and a pressure loading segment. During the movement of the locking head according to the three-segment motion speed curve, the temperature signal output by the temperature sensor array, the pressure signal output by the pressure sensor array, and the displacement signal output by the displacement sensor are synchronously collected at a preset sampling period. The preset sampling period is less than one-fifth of the dominant time constant of the locking head mechanical system. The validity of the synchronously acquired temperature signal, pressure signal and displacement signal is verified. The validity verification includes temperature value range verification, pressure value monotonicity verification and displacement-pressure consistency verification. Abnormal data that fails the validity check is removed, and the remaining valid data is stamped with a unified timestamp to generate temperature field data, pressure field data, and displacement data.

3. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 1, characterized in that, Based on the temperature field data, the temperature-mechanical property mapping database is queried to identify the current equivalent elastic modulus and equivalent damping coefficient of the smart buffer material, including: Spatial interpolation is performed on the discrete temperature values ​​of each temperature measurement point in the temperature field data to reconstruct the continuous two-dimensional temperature field distribution at the contact interface between the smart buffer material and the chip. Based on the continuous two-dimensional temperature field distribution, the area-weighted average temperature value of the contact interface is calculated using the area of ​​the influence region represented by each temperature measuring point as the weight, and is used as the characteristic temperature at the current moment. Using the characteristic temperature as the query index, table lookup interpolation is performed in the temperature-mechanical property mapping database to obtain the global equivalent elastic modulus and the global equivalent damping coefficient; Calculate the deviation between the temperature value of each temperature measuring point and the characteristic temperature. For measuring points whose deviation exceeds the preset gradient threshold, perform independent lookup table interpolation based on the local temperature of that measuring point to obtain the local mechanical state correction matrix.

4. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 3, characterized in that, Calculate the deviation between the temperature value at each temperature measuring point and the characteristic temperature. For measuring points where the deviation exceeds a preset gradient threshold, perform independent lookup table interpolation using the local temperature of that measuring point to obtain a local mechanical state correction matrix, including: Calculate the temperature deviation between the temperature value of each temperature measuring point and the characteristic temperature, and mark the temperature measuring points whose absolute value of the temperature deviation exceeds the preset gradient threshold as local correction measuring points. The local temperature of each of the local correction measurement points is used as the query index instead of the characteristic temperature. The table lookup and interpolation are performed in the temperature-mechanical property mapping database to obtain the local equivalent elastic modulus and local equivalent damping coefficient corresponding to each of the local correction measurement points. Based on the spatial position of each local correction measuring point on the contact interface, the local equivalent elastic modulus and the local equivalent damping coefficient are organized into a local mechanical state correction matrix. The elements at the non-local correction measuring point positions in the local mechanical state correction matrix are the global equivalent elastic modulus and the global equivalent damping coefficient.

5. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 4, characterized in that, Based on the current equivalent elastic modulus, equivalent damping coefficient, and displacement data, the pressure field data is dynamically compensated and corrected to generate an estimated value of the effective pressure distribution on the chip surface, including: The force transmission process of the smart buffer material is equivalent to a first-order inertial element, and the time constant of the first-order inertial element is determined by the ratio of the current equivalent damping coefficient to the current equivalent elastic modulus. For the pressure value sequence of each pressure measuring point in the pressure field data, dynamic compensation calculation is performed using the inverse model of the first-order inertial element. The dynamic compensation calculation is performed in the discrete time domain. The compensated pressure value at the current moment is determined by the original pressure value at the current moment, the compensated pressure value at the previous moment, the current equivalent elastic modulus, and the current equivalent damping coefficient. For the pressure measuring points corresponding to the spatial positions of each local correction measuring point in the local mechanical state correction matrix, the local equivalent elastic modulus and the local equivalent damping coefficient corresponding to the local correction measuring point are used to replace the current equivalent elastic modulus and the current equivalent damping coefficient, and the dynamic compensation calculation is performed independently. The compensated pressure values ​​of each pressure measuring point are arranged according to their spatial location to generate an estimated value of the effective pressure distribution on the chip surface.

6. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 5, characterized in that, For the pressure value sequence of each pressure measuring point in the pressure field data, dynamic compensation calculation is performed using the inverse model of the first-order inertial element, including: The inverse model of the first-order inertial element is discretized in the discrete time domain to obtain the compensation operation rules in the form of first-order difference equations. The compensation coefficient of the first-order difference equation is updated in real time based on the current equivalent elastic modulus and the current equivalent damping coefficient. The compensation coefficient is inversely proportional to the time constant of the first-order inertial element. For each pressure measuring point, the compensation calculation rule is executed cycle by cycle: The compensated pressure value of the current sampling period is equal to the weighted combination of the original pressure value of the current sampling period and the compensated pressure value of the previous sampling period. The weight of the weighted combination is determined by the compensation coefficient that is updated in real time. When the current equivalent elastic modulus increases, the compensation coefficient increases accordingly, thereby increasing the weight of the original pressure value in the current sampling period in the weighted combination. When the current equivalent damping coefficient increases, the compensation coefficient decreases accordingly, thereby increasing the weight of the compensated pressure value from the previous sampling period in the weighted combination.

7. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 1, characterized in that, Based on the estimated effective pressure distribution on the chip surface and combined with the equivalent stress safety threshold, the chip's safe pressure-bearing state is assessed, and a global safety risk level identifier is generated, including: The estimated effective pressure distribution on the chip surface is mapped to each monitoring micro-element on the chip surface. Spatial interpolation is performed on the monitoring micro-element that is not directly covered by the pressure measurement point to generate a continuous effective pressure field distribution covering the entire chip surface. Calculate the area-weighted average effective pressure value of the continuous effective pressure field distribution, compare the area-weighted average effective pressure value with the compressive strength threshold, and generate a uniform compressive stress evaluation result; Based on the spatial gradient of the continuous effective pressure field distribution, combined with the chip thickness and back support conditions, the bending stress of each monitoring micro-element is calculated according to the thin plate bending theory. The bending stress is then compared with the equivalent stress safety threshold to generate a bending stress assessment result. Based on the uniform compressive stress assessment results and the bending stress assessment results, a comprehensive determination of the global safety risk level is made. The global safety risk level includes three levels: safe state, early warning state, and dangerous state.

8. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 7, characterized in that, Based on the spatial gradient of the continuous effective pressure field distribution, and combined with the chip thickness and back support conditions, the bending stress of each monitoring micro-element is calculated according to the thin plate bending theory, including: For each of the monitoring micro-elements, the effective pressure values ​​of two monitoring micro-elements adjacent to the monitoring micro-element are taken along the first horizontal direction, and the first pressure gradient of the monitoring micro-element in the first horizontal direction is calculated by the central difference method. The effective pressure values ​​of two monitoring micro-elements adjacent to the monitoring micro-element are taken along the second horizontal direction, and the second pressure gradient of the monitoring micro-element in the second horizontal direction is calculated by the central difference method. The second horizontal direction is orthogonal to the first horizontal direction. Based on the first pressure gradient and the second pressure gradient, combined with the chip thickness and back support conditions, the first bending stress component and the second bending stress component of the monitoring micro-element in the first horizontal direction and in the second horizontal direction are calculated respectively according to the thin plate bending theory. The bending stress value of the monitoring micro-element is obtained by vector synthesis of the first bending stress component and the second bending stress component.

9. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 1, characterized in that, When the global security risk level is identified as a warning state, a preventative fine-tuning instruction is generated. This preventative fine-tuning instruction includes the target adjustment displacement and target adjustment speed of the locking head, including: The difference between the estimated effective pressure distribution on the chip surface and the target pressure distribution is calculated as the desired pressure adjustment amount; Based on the inverse solution operation of the multiphysics coupling calibration model of the locking system, combined with the current equivalent elastic modulus, the target adjustment displacement of the locking head is estimated according to the expected pressure adjustment amount. Based on the target adjustment displacement, an S-shaped velocity planning curve with continuously varying acceleration is used to generate the target adjustment speed of the locking head; Based on the rate of change of the current chip stress state, an adaptive response time limit is set to generate a preventive fine-tuning instruction that includes the target adjustment displacement, the target adjustment speed, and the response time limit.

10. The adaptive control method for locking force based on temperature-pressure dual feedback according to claim 2, characterized in that, The movement speed curve of the locking head is set according to the locking process parameters, and the locking process is started, including: Based on the specification identification code of the chip to be mounted, retrieve the corresponding latching process parameters from the process database; During the rapid approach phase, the locking head is controlled to move toward the chip mounting surface at a first speed until the gap between the end face of the locking head and the upper surface of the smart buffer material is reduced to a preset approach gap threshold. When the gap decreases to the preset near gap threshold, switch to the buffer contact section and control the locking head to continue descending at a second speed lower than the first speed until the pressure sensor array detects a pressure signal greater than the preset contact determination threshold. When the pressure sensor array detects a pressure signal greater than the preset contact determination threshold, it switches to the pressure loading section and controls the locking head to continue pressing down at a preset loading rate until the target pressure value or the target displacement value is reached, and the locking process is started.

Images