A Method and System for Selective Wave Soldering Repair of Short-Circuit Solder Joints in PCBA

Abstract

This invention relates to the field of electronic manufacturing technology, specifically a method and system for selective wave soldering repair of short-circuit solder joints in PCBAs. The method generates an anisotropic thermal inertia spectrum characterizing the anisotropy of solder thermal conductivity by acquiring infrared thermal images of short-circuit solder joints after wave soldering. This spectrum, combined with geometric features, is used to select suitable target short-circuit solder joints for repair. The direction of the thermodynamically weak interface in the bridging region is identified, and the nozzle trajectory and directional Lorentz force parameters are planned. While the nozzle travels along the trajectory and heats the solder to form locally molten solder, a directional electromagnetic field is applied to drive the molten solder to migrate directionally along the thermodynamically weak interface direction to break the short-circuit bridge. Upon detecting the bridge breakage, negative pressure backflow and termination of the electromagnetic field are executed. This invention achieves directional and controllable migration of the molten solder, avoiding problems such as secondary bridging caused by uncontrollable solder flow in traditional selective soldering repair, thus improving the accuracy and consistency of short-circuit solder joint repair.

Description

Technical Field

[0001] This invention relates to the field of electronic manufacturing technology, and in particular to a method and system for selectively repairing short-circuit solder joints in PCBA based on selective wave soldering. Background Technology

[0002] After wave soldering, PCBAs often experience short-circuit defects due to solder bridging caused by dense component leads, reduced pad spacing, and fluctuating flux activity. Traditional repair methods for these short-circuit solder joints rely primarily on manual soldering. Operators use soldering irons and desoldering wires to manually heat and remove solder from the bridging area. However, manual soldering is highly dependent on operator experience, resulting in poor consistency and low efficiency. Furthermore, on fine-pitch devices, it can easily damage adjacent components due to thermal shock or accidental contact. In recent years, the industry has attempted to apply selective wave soldering equipment to the automated repair of short-circuit solder joints. This involves targeted heating of the bridging area, supplemented by negative pressure suction or airflow purging to remove excess solder. However, existing selective wave soldering repair methods generally only screen repairable solder joints based on geometric dimensions and use fixed trajectories and constant energy parameters for heating and solder removal. They do not consider the anisotropic thermal conductivity characteristics determined by the solder solidification orientation and microstructure differences within the bridging region. This results in uncontrollable solder flow direction after heating and melting, and problems such as solder tailing, secondary bridging, or pad peeling are prone to occur during repair. In particular, when dealing with bridging with irregular geometric shrinkage morphology, the repair success rate drops significantly, which restricts the depth of automated application of selective wave soldering in the field of short circuit repair. Summary of the Invention

[0003] This invention overcomes the shortcomings of the prior art and provides a method and system for selectively repairing short-circuit solder joints in PCBA based on selective wave soldering.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The first aspect of this invention discloses a method for selectively repairing short-circuit solder joints in PCBAs based on selective wave soldering, comprising the following steps: S1. Obtain short-circuit solder joint images of PCBA after wave soldering and generate thermal inertia anisotropy maps characterizing the anisotropy of solder thermal conduction based on their infrared thermal imaging. S2. Based on the thermal inertia anisotropy map and the geometric morphology of the short-circuit solder joints, target short-circuit solder joints whose thermal inertia anisotropy in the bridging region meets the preset threshold condition are selected. S3. Identify the direction of the minimum thermal inertia principal axis of the bridging region of the target short-circuit weld point and determine it as the direction of the thermodynamic weak interface, and plan a repair strategy that includes the weld nozzle motion trajectory and directional Lorentz force parameters. S4. Control the welding nozzle to move along the welding nozzle trajectory and heat the bridging area to form locally molten solder. At the same time, apply an electromagnetic field corresponding to the directional Lorentz force parameter to drive the molten solder to migrate directionally along the thermodynamic weak interface direction and break the short-circuit bridge. S5. Real-time monitoring of physical quantities characterizing the bridging disconnection. When the bridging is detected to be disconnected, control the welding nozzle to perform negative pressure back suction and simultaneously terminate the application of the electromagnetic field.

[0005] Furthermore, S1 specifically refers to: Images of short-circuit solder joints in PCBA after wave soldering are simultaneously acquired using a visible light camera and an infrared thermal imager to obtain visible light images and time-series multi-frame infrared thermal images of the short-circuit solder joints. Based on the visible light image, the short-circuit bridging region is identified and segmented, and the segmented short-circuit bridging region is spatially mapped onto the multi-frame infrared thermal image to obtain the infrared thermal image sequence of the bridging region. The thermal response curves of temperature change over time for each pixel are extracted from the infrared thermal image sequence. Based on the slope and attenuation characteristics of the temperature rise segment of the thermal response curves, the thermal inertia scalar values ​​of each pixel are calculated, and a thermal inertia scalar distribution map of the bridging region is generated. In the thermal inertia scalar distribution map, the rate of change of thermal inertia is calculated for each pixel along different directions to obtain the thermal inertia direction gradient of each pixel, and the thermal inertia extreme value direction of each pixel is determined according to the thermal inertia direction gradient to construct the thermal inertia vector field of the bridging region. Spatial clustering is performed on the thermal inertia vector field, and the consistent principal axis of the thermal inertia extreme value direction within the clustering region is extracted. The thermal inertia principal axis direction and the thermal inertia anisotropy characterizing the degree of directional difference at each location in the bridging region are then fitted. The thermal inertia principal axis direction and thermal inertia anisotropy are mapped to the geometric space of the bridging region to generate a thermal inertia anisotropy map.

[0006] Furthermore, S2 specifically refers to: The thermal inertia anisotropy values ​​of each pixel in the bridging region are extracted from the thermal inertia anisotropy map. At the same time, the geometric features of the bridging region are extracted from the short-circuit solder joint image to obtain the width distribution of the bridging and the spacing between the solder pads at both ends of the bridging. Based on the width distribution of the bridge, identify the geometric contraction of the bridge extending longitudinally, and delineate the thermally induced separation potential zone at the geometric contraction. Within the thermally induced separation potential region, the regional average value of the thermal inertia anisotropy is calculated and compared with a preset lower limit threshold for anisotropy. When the average value of the region reaches or exceeds the preset lower limit threshold of anisotropy, the corresponding short-circuit solder joint is marked as a candidate repairable solder joint. For the candidate repairable solder joints, the directional deviation angle between the principal axis of thermal inertia in the thermally induced separation potential zone and the longitudinal direction of the bridging is obtained. When the directional deviation angle is less than the preset allowable deviation angle threshold, the corresponding candidate repairable solder joint is determined as the target short-circuit solder joint suitable for selective soldering repair.

[0007] Furthermore, S3 specifically refers to: From the thermal inertia anisotropy map of the target short-circuit weld point, the thermal inertia principal axis direction field and the corresponding anisotropy scalar field of the bridging region are extracted, and the anisotropy scalar field is spatially differentiated to construct an anisotropy gradient vector field that characterizes the degree and direction of anisotropy change. In the anisotropy gradient vector field, starting from the geometric contraction of the bridging region, streamline tracing is performed along the negative anisotropy gradient direction to generate a thermodynamic weak interface path extending from the geometric contraction to the pad safety boundary, such that the tangent direction at each point of the thermodynamic weak interface path corresponds to the principal axis direction of the local minimum thermal inertia. Extract the principal axis directions of thermal inertia at each point along the thermodynamic weak interface path, and select the principal axis direction of thermal inertia with the smallest angle to the tangent direction of the thermodynamic weak interface path as the local weak interface orientation vector at the corresponding point. Determine the ordered set of all local weak interface orientation vectors on the path as the thermodynamic weak interface direction. The thermodynamic weak interface path is spline smoothed, and an offset path is obtained by translating along the normal of the thermodynamic weak interface path according to the preset offset distance between the welding nozzle and the bridging region. The offset path is used as the welding nozzle motion trajectory. At each point along the thermodynamic weak interface path, based on the anisotropy value of each point and the curvature corresponding to the bias path, the current amplitude parameters and magnetic field direction parameters of the directional Lorentz force that match the local weak interface orientation vector of the corresponding point are determined, and synchronously correlated along the welding nozzle movement trajectory according to position to form the directional Lorentz force parameters.

[0008] Furthermore, S4 specifically refers to: Control the welding nozzle to move to the starting point of the welding nozzle movement trajectory, and preheat the surface of the bridging area with the preset preheating power so that the solder in the bridging area reaches the solid plastic temperature range and forms a solder softening zone. The welding nozzle moves from the starting point to the ending point along the welding nozzle movement trajectory. During the movement, the heating power of the welding nozzle is switched from preheating power to melting power. Local melting zones are formed point by point along the welding nozzle movement path within the solder softening zone. The leading edge of the molten pool in the local melting zone is aligned with the direction of the thermodynamic weak interface. At the synchronous moment of the welding nozzle's movement, the corresponding induced current amplitude parameter and magnetic field direction parameter are retrieved from the directional Lorentz force parameter according to the current position of the welding nozzle, and a transient excitation current corresponding to the induced current amplitude parameter is applied to the electrode pair set on the side of the bridging region. At the same time, the excitation coil built into the welding nozzle is excited to generate a directional transient magnetic field corresponding to the magnetic field direction parameter. The transient excitation current is conducted through the molten solder in the local melting zone to form an induced current. The induced current interacts with the directional transient magnetic field to generate a directional Lorentz force, which drives the molten solder in the local melting zone to migrate along the direction of the thermodynamic weak interface to the outside of the solder nozzle movement trajectory, forming a molten solder migration flow. As the welding nozzle continues to travel, the molten solder migration flow continuously delivers molten solder to the outside of the bridge along the direction of the thermodynamic weak interface, and the width of the bridge cross section continues to narrow until the bridge is completely broken at the geometric contraction part, forming the bridge break end face.

[0009] Furthermore, S5 specifically refers to: As the welding nozzle travels along its trajectory and applies a directional Lorentz force to the bridging region, a high-frequency, low-amplitude probe current signal superimposed on the transient excitation current is continuously applied to the electrode pair, and the response voltage signals at both ends of the electrode pair are acquired simultaneously. The AC impedance component with the same frequency as the high-frequency low-amplitude detection current signal is extracted from the response voltage signal. Based on the real-time change of the AC impedance component in the welding nozzle travel sequence, a dynamic impedance spectrum characterizing the connectivity state of the molten solder in the local melting zone is constructed. The impedance rise trend corresponding to the narrowing process of the bridge cross section in the dynamic impedance spectrum is monitored. When an instantaneous impedance step change is detected in the impedance rise trend and the amplitude of the instantaneous impedance step change exceeds the preset circuit breaker threshold, a bridge disconnection confirmation signal is generated. In response to the bridge disconnection confirmation signal, the transient excitation current and high-frequency low-amplitude detection current signal applied to the electrode pair are immediately turned off, and the negative pressure suction channel at the end of the welding nozzle is activated to form a directional suction flow field at the bridge disconnection end face, which sucks the free molten solder that has migrated to the vicinity of the welding nozzle into the inner cavity of the welding nozzle. A reverse decaying current is applied to the excitation coil built into the welding nozzle according to a preset demagnetization timing curve, so that the magnetic field strength generated by the excitation coil gradually decays from the current directional transient magnetic field level to zero within a predetermined demagnetization time, thus completing the termination of the electromagnetic field application.

[0010] The second aspect of the present invention discloses a PCBA short-circuit solder joint selective repair system based on selective wave soldering. The system includes a memory and a processor. The memory stores a program for a PCBA short-circuit solder joint selective repair method based on selective wave soldering. When the program for the PCBA short-circuit solder joint selective repair method based on selective wave soldering is executed by the processor, the steps of any of the methods described in the present invention are implemented.

[0011] This invention addresses the technical deficiencies in the prior art and offers the following advantages: By introducing a thermal inertia anisotropy map to characterize the internal thermal conductivity direction preference of the solder, and by screening suitable target solder joints for repair and planning repair paths along thermodynamically weak interfaces, the invention utilizes directional Lorentz force to drive the molten solder to migrate controllably along a predetermined direction to break bridging. This transforms the repair process from "experience-driven" to "physical property-driven," effectively overcoming problems such as tailing, secondary bridging, and pad peeling caused by uncontrollable solder flow direction in traditional selective soldering repair. It improves the success rate and consistency of repairing fine-pitch and irregularly bridging solder joints, reduces the risk of thermal shock to adjacent components during the repair process, and reduces the need for manual intervention, thereby improving the automation level and process reliability of PCBA short-circuit defect repair. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0013] Figure 1 Flowchart of a PCBA short-circuit solder joint selective repair method based on selective wave soldering; Figure 2 This is a system architecture diagram for selective wave soldering repair of PCBA short-circuit solder joints. Detailed Implementation

[0014] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0015] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0016] like Figure 1 As shown, the first aspect of this invention discloses a method for selectively repairing short-circuit solder joints in PCBA based on selective wave soldering, comprising the following steps: S1. Obtain short-circuit solder joint images of PCBA after wave soldering and generate thermal inertia anisotropy maps characterizing the anisotropy of solder thermal conduction based on their infrared thermal imaging. S2. Based on the thermal inertia anisotropy map and the geometric morphology of the short-circuit solder joints, target short-circuit solder joints whose thermal inertia anisotropy in the bridging region meets the preset threshold condition are selected. S3. Identify the direction of the minimum thermal inertia principal axis of the bridging region of the target short-circuit weld point and determine it as the direction of the thermodynamic weak interface, and plan a repair strategy that includes the weld nozzle motion trajectory and directional Lorentz force parameters. S4. Control the welding nozzle to move along the welding nozzle trajectory and heat the bridging area to form locally molten solder. At the same time, apply an electromagnetic field corresponding to the directional Lorentz force parameter to drive the molten solder to migrate directionally along the thermodynamic weak interface direction and break the short-circuit bridge. S5. Real-time monitoring of physical quantities characterizing the bridging disconnection. When the bridging is detected to be disconnected, control the welding nozzle to perform negative pressure back suction and simultaneously terminate the application of the electromagnetic field.

[0017] Specifically, S1 is: Images of short-circuit solder joints in PCBA after wave soldering are simultaneously acquired using a visible light camera and an infrared thermal imager to obtain visible light images and time-series multi-frame infrared thermal images of the short-circuit solder joints. Based on the visible light image, the short-circuit bridging region is identified and segmented, and the segmented short-circuit bridging region is spatially mapped onto the multi-frame infrared thermal image to obtain the infrared thermal image sequence of the bridging region. The thermal response curves of temperature change over time for each pixel are extracted from the infrared thermal image sequence. Based on the slope and attenuation characteristics of the temperature rise segment of the thermal response curves, the thermal inertia scalar values ​​of each pixel are calculated, and a thermal inertia scalar distribution map of the bridging region is generated. In the thermal inertia scalar distribution map, the rate of change of thermal inertia is calculated for each pixel along different directions to obtain the thermal inertia direction gradient of each pixel, and the thermal inertia extreme value direction of each pixel is determined according to the thermal inertia direction gradient to construct the thermal inertia vector field of the bridging region. Spatial clustering is performed on the thermal inertia vector field, and the consistent principal axis of the thermal inertia extreme value direction within the clustering region is extracted. The thermal inertia principal axis direction and the thermal inertia anisotropy characterizing the degree of directional difference at each location in the bridging region are then fitted. The thermal inertia principal axis direction and thermal inertia anisotropy are mapped to the geometric space of the bridging region to generate a thermal inertia anisotropy map.

[0018] It should be noted that for each pixel in the infrared thermographic sequence of the bridging region, its thermal response curve over time is extracted. This curve includes a temperature rise segment caused by residual heat from wave soldering or preheating, and a subsequent natural cooling attenuation segment. The calculation of the thermal inertia scalar value is based on the morphological characteristics of these two curves: the slope of the temperature rise segment reflects the response rate of the solder to heat input at that pixel; a larger slope indicates a more sensitive thermal response. The characteristics of the attenuation segment reflect the rate at which heat is retained or diffused outwards; a smoother attenuation indicates a higher thermal inertia. In practice, the temperature rise segment can be linearly fitted to obtain the temperature rise slope, and the time taken for the temperature to drop to half the difference between the initial temperature and the peak temperature in the attenuation segment can be extracted as the attenuation half-life. The product or weighted combination of the temperature rise slope and the attenuation half-life is used as the thermal inertia scalar value for that pixel. After calculating the thermal inertia scalar values ​​for all pixels in the bridging region, the thermal inertia scalar values ​​of each point are arranged according to pixel position, thus generating a thermal inertia scalar distribution map corresponding to the spatial resolution of the infrared thermographic image.

[0019] After obtaining the scalar distribution map of thermal inertia, a vector field of thermal inertia is further constructed to reveal the directional preference of heat conduction within the solder. Subsequently, spatial clustering is performed on the vector field to extract the overall heat conduction direction and directional consistency of local regions. Specifically, density-based spatial clustering or fixed grid partitioning can be used to divide the bridging region into several spatially continuous clusters. Statistical analysis is performed on the extreme thermal inertia directions of each pixel within each cluster. Principal component analysis is used to extract the first principal component direction of the extreme direction distribution within that region. This first principal component direction is the principal axis direction of thermal inertia for that cluster, representing the unified direction with the strongest thermal conductivity of the solder in the corresponding local region. Simultaneously, the angular deviation between the extreme direction of each pixel within the cluster and the obtained principal axis direction of thermal inertia is calculated. The statistical measure of the deviation (such as standard deviation or mean deviation angle) is used as the anisotropy of thermal inertia in that region; the smaller the deviation, the higher the anisotropy value, indicating a more significant directional preference of the solder along the principal axis direction for heat conduction in that region. Thus, each cluster region obtains a set of principal axes of thermal inertia and the corresponding degree of thermal inertia anisotropy.

[0020] Finally, a one-to-one correspondence is established between the principal axes of thermal inertia and the anisotropy of thermal inertia of each cluster region obtained from the above processing and the geometric spatial location of the bridging region. For example, the center coordinates or region outline of each cluster region are associated with the thermal conduction direction vector and anisotropy attribute of that region, so that each position on the bridging two-dimensional geometry carries the physical information of thermal conduction direction and anisotropy, thereby generating a complete thermal inertia anisotropy map.

[0021] Specifically, S2 is: The thermal inertia anisotropy values ​​of each pixel in the bridging region are extracted from the thermal inertia anisotropy map. At the same time, the geometric features of the bridging region are extracted from the short-circuit solder joint image to obtain the width distribution of the bridging and the spacing between the solder pads at both ends of the bridging. Based on the width distribution of the bridge, identify the geometric contraction of the bridge extending longitudinally, and delineate the thermally induced separation potential zone at the geometric contraction. Within the thermally induced separation potential region, the regional average value of the thermal inertia anisotropy is calculated and compared with a preset lower limit threshold for anisotropy. When the average value of the region reaches or exceeds the preset lower limit threshold of anisotropy, the corresponding short-circuit solder joint is marked as a candidate repairable solder joint. For the candidate repairable solder joints, the directional deviation angle between the principal axis of thermal inertia in the thermally induced separation potential zone and the longitudinal direction of the bridging is obtained. When the directional deviation angle is less than the preset allowable deviation angle threshold, the corresponding candidate repairable solder joint is determined as the target short-circuit solder joint suitable for selective soldering repair.

[0022] It should be noted that the core of the screening process lies in determining whether the bridging region possesses shrinkage characteristics in its geometric structure that are conducive to thermally induced disconnection, and whether the thermal conductivity direction preference within the solder in this region matches the geometric extension direction of the bridging. For identifying geometrically contracted sections, this embodiment obtains a width variation curve along the longitudinal centerline of the bridging. The curve is composed of the distance values ​​at various longitudinal positions along the boundaries of both sides of the bridging profile. By searching for local minimum intervals on the curve where the width value is less than the average width of the adjacent regions on both sides, the bridging segment corresponding to the interval can be identified as a geometrically contracted section. For example, when a bridging exhibits a width that gradually narrows from both ends towards the middle in its longitudinal extension, and the narrowest point's width drops to less than half the average width of the pad ends on both sides, this narrowest point and its adjacent transition section are identified as geometrically contracted sections. Since the geometrically contracted section has the least amount of solder and the weakest structural continuity, it is a preferred area for disconnection under heating and melting conditions. After identifying the geometric contraction, a closed region covering the geometric contraction and its neighborhood is defined as the thermally induced separation potential zone, extending a predetermined length (e.g., to twice the width of the geometric contraction or to the position where the width returns to the local average width) along the longitudinal direction of the bridge connection, with the geometric contraction as the center.

[0023] Within the thermally induced separation potential zone, the thermal inertia anisotropy values ​​of all pixels within the zone are collected and their arithmetic mean is calculated as the regional average of thermal inertia anisotropy for the corresponding region. This regional average is compared with a preset lower threshold for anisotropy. The lower threshold is pre-calibrated based on solder composition and welding process conditions, and is exemplarily set to 50% of the maximum measurable anisotropy value. For example, when the anisotropy dimension is normalized to the 0-1 range, the lower threshold is 0.5. When the regional average reaches or exceeds this lower threshold, it indicates that the bridging has a sufficiently significant thermal conductivity direction preference near the geometric contraction, enabling controllable directional migration of molten solder under directional Lorentz force, thus marking the short-circuit solder joint as a candidate repairable solder joint. Conversely, it indicates that the thermal conductivity directionality of the bridging solder is not significant, making it difficult to precisely control the solder flow direction during repair, and thus unsuitable for repair using this method.

[0024] For the marked candidate repairable solder joints, the matching relationship between the thermal conduction direction and the bridging geometric extension direction needs to be further examined. Specifically, the principal axis direction of thermal inertia of each pixel in the thermally induced separation potential area is obtained and its vector average value is calculated to obtain the representative principal axis direction of thermal inertia in this area. At the same time, the direction of the line connecting the centers of the pads at both ends of the bridging is taken as the longitudinal direction of the bridging, and the angle between the representative principal axis direction of thermal inertia and the longitudinal direction of the bridging is calculated as the directional deviation angle. This directional deviation angle is compared with a preset allowable deviation angle threshold, which can be set to 15 degrees for example. When the directional deviation angle is less than this allowable threshold, it indicates that the dominant thermal conduction direction of the bridging solder is roughly consistent with the bridging extension direction. When the directional Lorentz force drives the molten solder to migrate, the solder will smoothly detach along the longitudinal direction of the bridging and will not be pushed to adjacent pads due to directional deviation, causing secondary bridging. This candidate repairable solder joint is finally determined as the target short-circuit solder joint suitable for selective soldering repair.

[0025] Specifically, S3 is: From the thermal inertia anisotropy map of the target short-circuit weld point, the thermal inertia principal axis direction field and the corresponding anisotropy scalar field of the bridging region are extracted, and the anisotropy scalar field is spatially differentiated to construct an anisotropy gradient vector field that characterizes the degree and direction of anisotropy change. In the anisotropy gradient vector field, starting from the geometric contraction of the bridging region, streamline tracing is performed along the negative anisotropy gradient direction to generate a thermodynamic weak interface path extending from the geometric contraction to the pad safety boundary, such that the tangent direction at each point of the thermodynamic weak interface path corresponds to the principal axis direction of the local minimum thermal inertia. Extract the principal axis directions of thermal inertia at each point along the thermodynamic weak interface path, and select the principal axis direction of thermal inertia with the smallest angle to the tangent direction of the thermodynamic weak interface path as the local weak interface orientation vector at the corresponding point. Determine the ordered set of all local weak interface orientation vectors on the path as the thermodynamic weak interface direction. The thermodynamic weak interface path is spline smoothed, and an offset path is obtained by translating along the normal of the thermodynamic weak interface path according to the preset offset distance between the welding nozzle and the bridging region. The offset path is used as the welding nozzle motion trajectory. At each point along the thermodynamic weak interface path, based on the anisotropy value of each point and the curvature corresponding to the bias path, the current amplitude parameters and magnetic field direction parameters of the directional Lorentz force that match the local weak interface orientation vector of the corresponding point are determined, and synchronously correlated along the welding nozzle movement trajectory according to position to form the directional Lorentz force parameters.

[0026] It should be noted that in the anisotropy map, each pixel in the bridging region carries two types of information: the principal axis direction of thermal inertia and the anisotropy value. The anisotropy scalar field refers to the spatial distribution of the anisotropy values ​​of each pixel within the bridging region. The magnitude of each pixel's value reflects the strength of the solder's thermal conductivity orientation preference at that location; a higher value indicates that the solder's thermal conductivity along its principal axis is significantly better than in other directions, while a lower value indicates that the thermal conductivity at that location tends towards isotropy. The principal axis direction field records the vector distribution of the principal axis directions of thermal inertia for each pixel.

[0027] To extract directional change information from the anisotropy scalar field, spatial differentiation is performed on the anisotropy scalar field. Specifically, for each pixel within the bridging region, the rate of change of the anisotropy value is calculated along both the horizontal and vertical axes of the coordinate system. These two rates of change are then combined to form the anisotropy gradient vector for that pixel. The magnitude of the gradient vector characterizes the drastic change in anisotropy at that location, while its direction points to the direction of the fastest increase in anisotropy value. After performing the above calculations for all pixels, an anisotropy gradient vector field corresponding to the bridging region space is constructed. The negative gradient direction of the gradient vector field points to the direction of the fastest decrease in anisotropy value, which is the location where the thermal conductivity bias is weakest and the solder behavior is closest to isotropy.

[0028] Based on the anisotropy gradient vector field, starting from the geometric contraction of the bridging region, streamline tracing is performed along the negative anisotropy gradient direction to generate a thermodynamically weak interface path. The streamline tracing process can be analogized to releasing a virtual particle from the starting point, causing the particle to move along the negative gradient direction within the anisotropy gradient vector field; its trajectory forms a streamline. Since the negative gradient direction points to the region of lowest anisotropy, the streamline will sequentially pass through the locations with the weakest thermal conductivity bias within the bridging region during its extension. The principal axes of local thermal inertia corresponding to these locations are the principal axes of local minimum thermal inertia at those locations. Therefore, the tangent direction at each point of the streamline naturally corresponds to the principal axis of local minimum thermal inertia. Streamline tracing continues until the pad safety boundary is reached. This pad safety boundary is a preset virtual limit that maintains a safe distance from adjacent normal pads; for example, it can be set as an envelope at least 0.5 mm from the edge of the adjacent pad. The resulting streamlines are the thermodynamic weak interface paths extending from the geometric contraction to the pad safety boundary, representing the spatial orientation of the bridging region where thermal conductivity is weakest and the solder is most thermodynamically driven to migrate in a directional manner by external forces.

[0029] Based on this, the principal axes of thermal inertia recorded at each point along the aforementioned thermodynamic weak interface path are processed. The principal axis of thermal inertia with the smallest angle to the tangent direction of the path is selected as the local weak interface orientation vector at that point. The ordered set of all local weak interface orientation vectors along the path constitutes the thermodynamic weak interface direction. Subsequently, spline smoothing is performed on the thermodynamic weak interface path to eliminate local jitter during the tracking process. Based on the preset offset distance between the welding nozzle and the bridging area surface, the smoothed path is translated along its normal direction to obtain an offset path that maintains a constant distance from the bridging surface. This offset path is the welding nozzle motion trajectory. Finally, along this welding nozzle motion trajectory, the directional Lorentz force current amplitude parameters and magnetic field direction parameters that match the local weak interface orientation vector at each point are determined according to their positions, forming directional Lorentz force parameters synchronously associated with the welding nozzle motion trajectory for use in subsequent repair execution stages.

[0030] Specifically, based on the anisotropy values ​​at each point and the curvature corresponding to the bias path, the current amplitude parameters and magnetic field direction parameters of the directional Lorentz force that match the local weak interface orientation vector at the corresponding point are determined as follows: Extract each sampling point along the thermodynamic weak interface path at a preset interval, and obtain the local weak interface orientation vector, anisotropy value, radius of curvature at the corresponding position on the bias path corresponding to the sampling point, and preset welding nozzle travel speed at each sampling point. The local weak interface orientation vector at each sampling point is determined as the direction of the target driving force of the molten solder. A local flow coordinate system that varies with the sampling point is established with the direction of the target driving force of the molten solder as the vertical axis and the direction perpendicular to the direction of the target driving force of the molten solder in the pad plane as the horizontal axis. Within the local flow coordinate system, based on the anisotropy value at the sampling point, the thermodynamic migration resistance coefficient at the sampling point is obtained through a preset mapping relationship between anisotropy and thermodynamic migration resistance coefficient; and based on the radius of curvature and nozzle travel speed at the corresponding position on the offset path, the centrifugal induced offset force coefficient generated at the sampling point due to the bending of the nozzle's motion trajectory is calculated. The thermodynamic migration resistance coefficient and the centrifugal induced offset force coefficient are vector summed to obtain the total resistance vector of solder migration at the sampling point. Based on the conductivity parameter of the solder in the molten state and the preset nominal value of the applied magnetic induction intensity, the amplitude parameter of the induced current required to generate the Lorentz force along the direction of the target driving force of the molten solder and with an amplitude equal to the amplitude of the total resistance vector of solder migration is determined. Based on the chirality rule between the direction of the Lorentz force, the direction of the induced current and the direction of the magnetic field, the magnetic field direction parameters that match the amplitude parameters of the induced current at the corresponding sampling point and the direction of the target driving force of the molten solder are determined in the local flow coordinate system, so that the generated directional Lorentz force coincides with the direction of the target driving force of the molten solder. The induced current amplitude parameters and magnetic field direction parameters of each sampling point are matched one-to-one with the position coordinates of the bias path to form directional Lorentz force parameters that are synchronously associated with the welding nozzle movement trajectory.

[0031] It should be noted that a series of sampling points are extracted along the thermodynamically weak interface path at preset intervals, and the local weak interface orientation vector, anisotropy value, radius of curvature at the corresponding position on the offset path, and preset solder nozzle travel speed are obtained at each sampling point. To facilitate independent analysis of force balance at each sampling point, the direction pointed to by the local weak interface orientation vector at that point is taken as the target driving force direction of the molten solder. A local flow coordinate system is established with this direction as the vertical axis and the direction perpendicular to this direction in the solder pad plane as the horizontal axis, which changes with each sampling point. In the local flow coordinate system, the resistance to solder migration mainly comes from two aspects. One is the thermodynamic migration resistance caused by the anisotropy of solder thermal conductivity, the magnitude of which is related to the anisotropy value at that point. The mapping relationship between anisotropy and thermodynamic migration resistance coefficient is established in advance through experimental calibration. For example, a lookup table method or piecewise linear mapping can be used to map the anisotropy value to the corresponding resistance coefficient. A high anisotropy value at a sampling point indicates a greater microstructural resistance encountered by the solder migrating along the non-principal axis, resulting in a larger thermodynamic migration resistance coefficient. Conversely, at sampling points with lower anisotropy values, the solder tends to be isotropic, leading to a smaller migration resistance coefficient. The thermodynamic migration resistance coefficient at each sampling point is obtained in this way, with its direction along the horizontal axis of the local flow coordinate system, perpendicular to the target driving force direction. Another source of resistance is the centrifugal effect generated when the nozzle moves along a curved trajectory. Since the nozzle travels along an offset path, when this path bends at a point, the locally molten solder experiences centrifugal force due to the change in nozzle movement direction. The centrifugal-induced offset force coefficient is determined based on the radius of curvature and nozzle speed at the corresponding offset path position. A smaller radius of curvature and higher speed result in a more significant centrifugal-induced effect and a larger coefficient. The direction of the centrifugal-induced offset force coefficient is also along the horizontal axis of the local flow coordinate system, superimposed in the same direction as the thermodynamic migration resistance coefficient.

[0032] The total resistance vector for solder migration at the sampling point is obtained by vector summing the thermodynamic migration resistance coefficient and the centrifugal induced displacement force coefficient. This vector represents the magnitude of the lateral resistance that needs to be overcome to drive the molten solder to migrate along the target driving force direction at that sampling point. To generate a Lorentz force that precisely counteracts this total resistance and drives the molten solder to migrate along the target driving force direction, the required induced current amplitude parameter is determined based on the conductivity parameter of the molten solder and the preset nominal value of the applied magnetic induction intensity. Specifically, given the conductivity of the molten solder and the preset nominal value of the applied magnetic induction intensity in the solder region, the magnitude of the Lorentz force is determined by the product of the induced current amplitude and the nominal value of the magnetic induction intensity. By setting the component of this Lorentz force perpendicular to the target driving force direction equal to the amplitude of the total resistance vector for solder migration, the required induced current amplitude parameter can be calculated.

[0033] After determining the amplitude parameters of the induced current, it is necessary to further determine the direction parameters of the magnetic field to ensure that the direction of the Lorentz force coincides with the direction of the target driving force. Based on the chirality rule among the Lorentz force direction, the induced current flow direction, and the magnetic field direction, in the local flow coordinate system, given that the induced current flows through the molten solder via the electrode pair in a defined direction, and with the required Lorentz force direction and induced current flow direction already determined, the required magnetic field direction parameters can be uniquely determined. These magnetic field direction parameters are used to control the excitation direction of the excitation coil built into the welding nozzle, ensuring that the directional transient magnetic field generated by the excitation coil, under the combined action of the induced current and the magnetic field, generates a directional Lorentz force whose direction precisely coincides with the target driving force direction of the molten solder.

[0034] The above process is executed sequentially at each sampling point along the thermodynamic weak interface path, and the induced current amplitude parameters and magnetic field direction parameters obtained at each sampling point are bound one by one with the position coordinates of the bias path. Finally, directional Lorentz force parameters synchronously associated with the position along the welding nozzle movement trajectory are formed, which can be retrieved in real time during the subsequent repair execution stage.

[0035] Specifically, S4 is: Control the welding nozzle to move to the starting point of the welding nozzle movement trajectory, and preheat the surface of the bridging area with the preset preheating power so that the solder in the bridging area reaches the solid plastic temperature range and forms a solder softening zone. The welding nozzle moves from the starting point to the ending point along the welding nozzle movement trajectory. During the movement, the heating power of the welding nozzle is switched from preheating power to melting power. Local melting zones are formed point by point along the welding nozzle movement path within the solder softening zone. The leading edge of the molten pool in the local melting zone is aligned with the direction of the thermodynamic weak interface. At the synchronous moment of the welding nozzle's movement, the corresponding induced current amplitude parameter and magnetic field direction parameter are retrieved from the directional Lorentz force parameter according to the current position of the welding nozzle, and a transient excitation current corresponding to the induced current amplitude parameter is applied to the electrode pair set on the side of the bridging region. At the same time, the excitation coil built into the welding nozzle is excited to generate a directional transient magnetic field corresponding to the magnetic field direction parameter. The transient excitation current is conducted through the molten solder in the local melting zone to form an induced current. The induced current interacts with the directional transient magnetic field to generate a directional Lorentz force, which drives the molten solder in the local melting zone to migrate along the direction of the thermodynamic weak interface to the outside of the solder nozzle movement trajectory, forming a molten solder migration flow. As the welding nozzle continues to travel, the molten solder migration flow continuously delivers molten solder to the outside of the bridge along the direction of the thermodynamic weak interface, and the width of the bridge cross section continues to narrow until the bridge is completely broken at the geometric contraction part, forming the bridge break end face.

[0036] It should be noted that the directional disconnection of short-circuit bridging is achieved through the coordinated control of nozzle movement, heating power switching, and electromagnetic field application. Specifically, in the initial stage of repair execution, the nozzle first moves to the starting point of its movement trajectory. At this time, the nozzle has not yet entered the melting heating state, but preheats the surface of the bridging area with a preset preheating power. The preheating power is selected to raise the temperature of the solder in the bridging area to the solid-plastic temperature range, which is above the glass transition temperature of the solder but below the solidus temperature. In this range, although the solder has not melted, it has softened significantly, and the inter-lattice binding force has weakened, providing thermodynamic preconditions for the subsequent formation of a local melting zone. After preheating, a continuous solder softening band is formed on the surface of the bridging area corresponding to the nozzle movement trajectory. Subsequently, the nozzle moves from the starting point to the ending point along the nozzle movement trajectory. In the initial stage of the movement, the nozzle heating power is switched from preheating power to melting power, which is sufficient to raise the temperature of the solder above the liquidus temperature to form a molten state. As the welding nozzle moves along the pre-formed solder softening zone, the solder within the softening zone responds more rapidly to heat input, and the formation of localized melting zones is more controllable. During the journey, localized melting zones gradually form at the points where the tip of the welding nozzle contacts the solder softening zone, and the direction of the molten pool front is consistent with the direction of the thermodynamic weak interface. This is ensured by the normal offset relationship between the welding nozzle's trajectory and the thermodynamic weak interface path.

[0037] As the welding nozzle moves, the induced current amplitude and magnetic field direction parameters corresponding to its current position are retrieved in real time from the directional Lorentz force parameters. A transient excitation current is applied to the electrode pair pre-positioned on the side of the bridging region, the amplitude of which is determined by the induced current amplitude parameters at the current position. The electrode pair is positioned such that the excitation current flows through the molten solder in the local melting zone, thereby forming an induced current path in the molten solder. Simultaneously, the excitation coil built into the welding nozzle is excited to generate a directional transient magnetic field, the spatial orientation of which is determined by the magnetic field direction parameters at the current position. The induced current interacts with the directional transient magnetic field, generating a directional Lorentz force within the molten solder, the direction of which coincides with the direction of the target driving force of the molten solder at the current position.

[0038] Driven by the directional Lorentz force, the molten solder in the locally molten zone overcomes the thermodynamic migration resistance and centrifugal-induced displacement resistance, continuously migrating along the direction of the thermodynamically weak interface towards the outside of the nozzle's movement trajectory, forming a molten solder migration flow. As the nozzle continues to move along its trajectory, new locally molten zones are constantly generated, and the molten solder migration flow continuously delivers solder material to the outside of the bridge. This process causes the bridge cross-sectional width to gradually narrow along the direction of the thermodynamically weak interface, and the cross-sectional area to continuously decrease. Since the geometric contraction section itself has the narrowest cross-section and the least amount of solder, when the nozzle reaches the vicinity of this point, the cross-sectional narrowing effect reaches a critical state, and the bridge completely breaks at the geometric contraction section, forming a bridge break end face. At this point, the short-circuit bridge is successfully separated.

[0039] Specifically, S5 is: As the welding nozzle travels along its trajectory and applies a directional Lorentz force to the bridging region, a high-frequency, low-amplitude probe current signal superimposed on the transient excitation current is continuously applied to the electrode pair, and the response voltage signals at both ends of the electrode pair are acquired simultaneously. The AC impedance component with the same frequency as the high-frequency low-amplitude detection current signal is extracted from the response voltage signal. Based on the real-time change of the AC impedance component in the welding nozzle travel sequence, a dynamic impedance spectrum characterizing the connectivity state of the molten solder in the local melting zone is constructed. The impedance rise trend corresponding to the narrowing process of the bridge cross section in the dynamic impedance spectrum is monitored. When an instantaneous impedance step change is detected in the impedance rise trend and the amplitude of the instantaneous impedance step change exceeds the preset circuit breaker threshold, a bridge disconnection confirmation signal is generated. In response to the bridge disconnection confirmation signal, the transient excitation current and high-frequency low-amplitude detection current signal applied to the electrode pair are immediately turned off, and the negative pressure suction channel at the end of the welding nozzle is activated to form a directional suction flow field at the bridge disconnection end face, which sucks the free molten solder that has migrated to the vicinity of the welding nozzle into the inner cavity of the welding nozzle. A reverse decaying current is applied to the excitation coil built into the welding nozzle according to a preset demagnetization timing curve, so that the magnetic field strength generated by the excitation coil gradually decays from the current directional transient magnetic field level to zero within a predetermined demagnetization time, thus completing the termination of the electromagnetic field application.

[0040] It should be noted that during the repair process, the system needs to determine in real time whether the bridging has been completely broken, and immediately terminate the electromagnetic field and recover the residual solder at the moment of breakage to prevent uncontrollable flow of molten solder without external constraint. This implementation achieves accurate sensing of bridging breakage events by superimposing a detection signal on the driving current and monitoring its impedance response. To this end, throughout the entire process of the soldering nozzle moving along the trajectory and continuously applying a directional Lorentz force to the bridging region, a high-frequency, low-amplitude detection current signal is always superimposed on the transient excitation current applied to the electrode pair on the side of the bridging region. The high-frequency, low-amplitude detection current signal refers to a weak AC signal with a frequency significantly higher than the transient excitation current and an amplitude much smaller than the driving current amplitude. Its frequency can be selected between kilohertz and megahertz, and its amplitude is at the milliampere level. It is used to sense the impedance change of the current path in real time without interfering with the migration behavior of the molten solder. The response voltage signal at both ends of the electrode pair is acquired simultaneously. The response voltage includes both the low-frequency voltage drop generated by the driving current and the high-frequency voltage drop component induced by the high-frequency detection current.

[0041] To extract information related to the bridging connectivity from the response voltage signal, phase-sensitive detection or quadrature demodulation is used to extract the AC impedance component with the same frequency as the high-frequency, low-amplitude probe current signal. For example, after passing the response voltage signal through a bandpass filter whose center frequency is aligned with the probe current frequency, it is phase-sensitively multiplied and low-pass filtered with the reference signal of the probe current to obtain the component reflecting the real part of the complex impedance of the molten solder current path. The magnitude of the real component is directly related to the conductive cross-sectional area and length of the molten solder bridging path between the electrode pairs: when the bridging cross-section is intact, the impedance is low; as the cross-section narrows, the impedance gradually increases; once the bridging is broken, the current path is interrupted, and the impedance will undergo a step change. Based on the real-time changes of the AC impedance component during the nozzle travel sequence, a dynamic impedance spectrum characterizing the molten solder connectivity in the local melting zone is constructed, i.e., the impedance value changes with the nozzle position or time.

[0042] The dynamic impedance spectrum is monitored. When the solder nozzle approaches the geometric contraction zone, the bridging cross-section width rapidly narrows, and the dynamic impedance spectrum shows a continuous upward trend. When a discontinuous, instantaneous, large jump in impedance value is detected, and the amplitude of this step change exceeds a preset circuit breaker threshold, the bridging is confirmed to be completely broken. For example, the AC impedance component of molten solder in the conducting state is typically on the order of several ohms. The circuit breaker threshold can be set to an instantaneous increase in impedance exceeding 10 ohms, or to the ratio of the current impedance value to the impedance value of the previous sampling period exceeding twice. Once the above conditions are met, a bridging break confirmation signal is generated.

[0043] In response to the disconnection confirmation signal, the system immediately shuts off the transient excitation current and high-frequency low-amplitude probe current signal applied to the electrode pair, stopping all electrical excitation; at the same time, it activates the negative pressure suction channel at the end of the solder nozzle, forming a directional suction flow field at the bridging disconnection end face, rapidly sucking the free molten solder that has migrated to the vicinity of the solder nozzle but has not completely detached from the molten solder migration flow into the inner cavity of the solder nozzle, preventing it from cooling at the disconnection end face to form solder residue or solder balls.

[0044] Regarding the electromagnetic field shutdown, to avoid the instantaneous reverse high voltage and magnetic field abrupt changes that would induce interference electromotive force in adjacent circuits due to directly cutting off the excitation coil current, a reverse decaying current is applied to the excitation coil according to a preset demagnetization timing curve. The preset demagnetization timing curve is a pre-defined command curve showing the current changing over time. It specifies how the excitation coil current gradually decays from its current level to zero, for example, using an exponential decay curve or a linear decay curve with a constant slope. Typical demagnetization times can be set from a few milliseconds to tens of milliseconds. Driven by this reverse decaying current, the magnetic field strength generated by the excitation coil gradually and smoothly decays from the current directional transient magnetic field level to zero, completing the termination of the electromagnetic field application. The entire process does not generate abrupt magnetic field changes, ensuring electrical safety and a smooth transition of the thermodynamic state at the end of the repair process.

[0045] In this embodiment, it also includes: After completing the selective soldering repair and electromagnetic field termination of all identified target short-circuited solder joints, obtain the thermal inertia scalar distribution map of all unrepaired solder joints on the same PCBA, excluding the repaired solder joints. Extract the spatial variation gradient of thermal inertia amplitude along the radial and tangential directions of the pad from the thermal inertia scalar distribution map of each unrepaired solder joint. Decompose the gradient of each pixel into irrotational gradient component and helical gradient component. Statistically analyze the directional curl distribution of the helical gradient component in different annular regions of the pad. Construct a thermal inertia helicity annular spectrum feature vector that characterizes the degree of thermal flow asymmetry of the unrepaired solder joint. The thermal inertia helicity ring spectrum feature vectors of all unrepaired solder joints are combined into a solder joint helical feature set. The thermal diffusion time cost on the heat conduction path between solder joints is used as the constraint distance metric. Thermal diffusion distance-preserving manifold dimensionality reduction is performed on the solder joint helical feature set to obtain the thermal diffusion embedding coordinates of each solder joint in the thermal diffusion manifold space. In the thermal diffusion manifold space, with the thermal diffusion embedding coordinate of each solder joint as the center, the solder joints in the nearest neighbor domain of its thermal diffusion distance are searched. The cumulative difference distance of helicity from the center solder joint to each nearest solder joint along the thermal diffusion path of the manifold surface is calculated. The ratio of the cumulative difference distance of helicity to the anisotropy of the thermal inertia of the center solder joint itself is used as the bridging thermal deflection fragility index of the corresponding solder joint. Spatially interpolate the bridging thermal deflection vulnerability index of all solder joints on the thermal diffusion embedding coordinate to form a continuous thermal vulnerability potential energy surface. Locate local potential energy ridges on the potential energy surface where the vulnerability index decreases from the center to the periphery. Connect the interconnected ridge segments with potential energy values ​​exceeding the critical vulnerability level to form a network of implicit bridging tendency vulnerability ridges. By using the thermal diffusion distance reconstruction kernel, each node on the implicit bridging tendency fragile ridge network is mapped point by point from the thermal diffusion manifold space back to the PCBA physical coordinate plane, and a predetermined thermodynamic influence radius is expanded outward along the mapped node path to form a thermodynamically fragile envelope boundary that delineates all implicit bridging incubation regions. This boundary is used to guide subsequent selective thermodynamic hardening.

[0046] It should be noted that after completing the selective repair and electromagnetic field termination of all identified short-circuited solder joints on the current PCBA, the thermal inertia scalar distribution map of all unrepaired solder joints not included in the repair scope is obtained. For each unrepaired solder joint, the spatial variation gradient of the thermal inertia amplitude along the radial and tangential directions of the pad is extracted from its thermal inertia scalar distribution map, and the gradient vector of each pixel is decomposed into irrotational gradient components and helical gradient components. The irrotational gradient component corresponds to the divergence part of the scalar field, reflecting the uniform change of thermal inertia along the radial direction; the helical gradient component corresponds to the curl part of the scalar field, reflecting the rotational asymmetry of thermal inertia variation. By statistically analyzing the directional curl distribution of the helical gradient components in different annular regions of the pad, a thermal inertia helical annular spectrum feature vector characterizing the degree of thermal flow circulation asymmetry of the unrepaired solder joint is constructed. This feature vector describes the rotational skewness of the heat conduction direction inside the solder joint in each annular region.

[0047] To reveal the intrinsic correlation between different solder joints along the heat conduction path, a heat diffusion time cost along the heat conduction path between solder joints is used as a constraint distance metric. A heat diffusion distance-preserving manifold dimensionality reduction is performed on the solder joint spiral feature set. This dimensionality reduction process constructs a heat diffusion time cost matrix and embeds the high-dimensional spiral feature set into a low-dimensional heat diffusion manifold space with the goal of keeping the heat diffusion time distance between each pair of solder joints as constant as possible. This yields the heat diffusion embedding coordinates of each solder joint in the heat diffusion manifold space. The Euclidean distance between the embedded coordinates approximately reflects the time cost required for thermal disturbances to propagate along the heat conduction path of the PCBA substrate between solder joints.

[0048] After obtaining the thermal diffusion embedding coordinates of each solder joint, a thermally induced vulnerability potential energy surface is further constructed based on the aforementioned bridging thermal deflection vulnerability index. Local potential energy ridges with vulnerability indices decreasing from the center to the periphery are located on the potential energy surface. The interconnected ridge segments with potential energy values ​​exceeding the critical vulnerability level are connected to form a network of implicit bridging tendency vulnerability ridges. This reveals the regional connectivity relationship on the PCBA where no actual bridging has occurred but the thermodynamic conditions already satisfy the bridging tendency.

[0049] To apply the above analysis results to actual physical reinforcement operations, a thermal diffusion distance reconstruction kernel is used to inversely map each node on the implicit bridging-oriented vulnerable ridge network from the thermal diffusion manifold space back to the PCBA physical coordinate plane. The inverse mapping is based on the correspondence between the thermal diffusion time cost distance in the original physical space and the manifold space, and reconstructs the corresponding coordinate positions of each node on the physical plane through a kernel function. For example, for a node in the manifold space, its corresponding coordinates on the PCBA physical plane can be determined by solving the weighted back projection of the node and each unrepaired solder joint in the manifold space using a weighted centering method. A predetermined thermodynamic influence radius is extended outward along the mapped node path, for example, set to 0.3 to 0.5 times the average spacing between adjacent solder joints based on the solder joint spacing and thermal diffusion coefficient, forming a thermodynamic vulnerable envelope boundary that delineates all implicit bridging incubation regions. This envelope boundary provides a quantitative basis for PCBA process optimization, enabling targeted preventive thermodynamic reinforcement of the implicit bridging incubation area, thereby reducing the risk of secondary bridging in subsequent wave soldering or rework processes and improving the soldering reliability of PCBA in long-term service.

[0050] In this embodiment, it also includes: After completing the selective soldering repair of the target short-circuited solder joint, for all functional solder joints on the same PCBA that have not undergone repair, the real part increment of the high-frequency AC impedance before and after repair is collected as the electrical performance change value. The mean value of thermal inertia anisotropy of the central region of the pad is extracted from the previous thermal inertia anisotropy spectrum of each functional solder joint. The electrical performance change value of each solder joint is paired with its mean value of thermal inertia anisotropy to form the corresponding electro-thermal response characteristic pair of the solder joint. Using all functional solder joints as nodes, the absolute value of the difference in electrical performance change values ​​between any two solder joints is calculated to form an electrical difference matrix; at the same time, the absolute value of the difference in the mean value of thermal inertia anisotropy between any two solder joints is calculated to form a thermal difference matrix; the electrical difference matrix and the thermal difference matrix are multiplied element by element to generate an initial electrical-thermal mismatch correlation matrix, the elements of which characterize the degree of difference in electrical performance response between solder joints caused by thermodynamic asymmetry. The electro-thermal mismatch correlation matrix is ​​symmetricized to separate the pure symmetric part and the pure antisymmetric part. The pure antisymmetric part is taken as the electro-thermal mismatch antisymmetric tensor. The non-zero elements in the electro-thermal mismatch antisymmetric tensor indicate that there is a directional electro-thermal response imbalance between the corresponding solder joint pairs, that is, the implicit repair-induced damage transmission relationship. The electro-thermal heterogeneous antisymmetric tensor is subjected to eigenvalue decomposition, and all its non-zero real eigenvalues ​​and their corresponding antisymmetric eigenmodes are extracted. The eigenvectors of each antisymmetric eigenmode are projected onto the physical coordinates of the corresponding solder joint to obtain the polarity component of the solder joint's tendency to dummy solder joint in each antisymmetric eigenmode. Select the dominant antisymmetric intrinsic modulus with the largest modulus length, and spatially map the polarity components of the solder joints' solder joint tendency according to the actual position of the solder joints on the PCBA. Draw a polarity boundary map of solder joint risk characterized by the boundary between positive and negative polarities. The area where the positive and negative components transition in the map is the location where the potential solder joint risk induced by repair accumulates.

[0051] It should be noted that during the directional Lorentz force-driven current repair of the target short-circuited solder joint, although local melting and electromagnetic field effects are completed under controlled conditions, the thermo-electric coupling effect between the repaired area and surrounding functional solder joints may still introduce microstructural changes within adjacent solder joints. These changes may not directly manifest as detectable open or short circuit defects, but may lie dormant as incomplete soldering or reduced solder joint fatigue life. Due to the differences in the anisotropy of thermal conduction of each functional solder joint, the thermal shock and induced current disturbances brought about by the repair process will produce inconsistent electrical performance responses between different solder joints. It is difficult to distinguish the spatial distribution pattern of this inconsistency by simply relying on the absolute change of electrical parameters of a single solder joint.

[0052] To address the aforementioned issues, this embodiment, after completing the selective soldering repair of all target short-circuited solder joints, collects the real part increment of the high-frequency AC impedance before and after repair for each functional solder joint on the same PCBA that has not undergone repair. This is used as an electrical performance change value characterizing the degree of impact of the repair process on the electrical performance of the solder joint. Simultaneously, the mean value of thermal inertia anisotropy in the central region of the pad is extracted from the previously constructed thermal inertia anisotropy map of each functional solder joint, reflecting the strength of the thermal conduction direction preference of that solder joint. The electrical performance change value of each solder joint is paired with its mean value of thermal inertia anisotropy to form an electro-thermal response characteristic pair, thereby linking the post-repair electrical performance drift with the inherent thermodynamic properties of the solder joint. Based on this, using all functional solder joints as nodes, the absolute value of the difference in electrical performance change values ​​between any two solder joints is calculated to construct an electrical difference matrix, and the absolute value of the difference in the mean value of thermal inertia anisotropy between any two solder joints is calculated to construct a thermal difference matrix. The electrical difference matrix and the thermal difference matrix are multiplied element-wise to generate an initial electrical-thermal mismatch correlation matrix, whose elements characterize the degree of difference in electrical performance response between corresponding solder joint pairs due to thermodynamic asymmetry.

[0053] To extract directional damage propagation information, the electro-thermal mismatch correlation matrix is ​​symmetricized. Adding the original matrix to its transpose and taking half yields a purely symmetric part, reflecting the bidirectional equilibrium response difference between solder joint pairs. Subtracting the original matrix from its transpose and taking half yields a purely antisymmetric part, the electro-thermal mismatch antisymmetric tensor. The antisymmetric tensor has significant physical meaning: its non-zero elements indicate a directional electro-thermal response imbalance between corresponding solder joint pairs, meaning that the electrical performance changes of one solder joint relative to the other do not follow the same trend. This directional characteristic is the mathematical representation of repair-induced damage propagation along a specific path. For example, if the antisymmetric tensor elements corresponding to solder joint A to solder joint B are positive, and the elements corresponding to solder joint B to solder joint A are equally negative, it indicates a unidirectional electro-thermal response imbalance in the direction from A to B, implying the path direction of the repair thermal-electric disturbance propagation from A to B.

[0054] To resolve the principal mode of this directional transmission, eigenvalue decomposition is performed on the electro-thermal heterosymmetric tensor. Since the eigenvalues ​​of an antisymmetric real matrix must be purely imaginary or zero, and appear in pairs of conjugates, in practice, the paired eigenvectors corresponding to all non-zero real eigenvalues ​​are extracted and combined to form antisymmetric eigenmodes. The components of the eigenvectors of each antisymmetric eigenmode are projected onto the physical coordinates of the corresponding solder joint. Each solder joint then obtains the polarity component of its virtual solder tendency in each antisymmetric eigenmode. The sign of this polarity component indicates the positional polarity of the solder joint in the damage transmission mode represented by the corresponding eigenmode. For example, a positive component indicates that the solder joint has a "source" tendency for damage transmission in the eigenmode, while a negative component indicates a "sink" tendency.

[0055] The dominant antisymmetric eigenmode with the largest modulus length is selected, and the polarity components of the solder joint's poor soldering tendency are spatially mapped according to the actual position of the solder joint on the PCBA. On the physical coordinate plane, a polarity boundary map of poor soldering risk is plotted with the gradient transition zone between positive and negative components as the region of interest. The zero or low value regions of the transition between positive and negative components in the map represent the locations where repair-induced potential poor soldering risks accumulate. These locations correspond to the interfaces where the electro-thermal response spatially reverses direction, and are the potential failure areas where microscopic damage accumulation within the solder joint is most concentrated. Therefore, this embodiment can perform systematic risk screening of unrepaired solder joints after selective soldering repair, providing a basis for re-inspection of solder joint reliability or local reinforcement.

[0056] like Figure 2 As shown, the second aspect of the present invention discloses a PCBA short-circuit solder joint selective repair system based on selective wave soldering. The system includes a memory and a processor. The memory stores a program for a PCBA short-circuit solder joint selective repair method based on selective wave soldering. When the program for the PCBA short-circuit solder joint selective repair method based on selective wave soldering is executed by the processor, the steps of any of the methods described in the present invention are implemented.

[0057] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for selectively repairing short-circuit solder joints in PCBA based on selective wave soldering, characterized in that, Includes the following steps: S1. Obtain short-circuit solder joint images of PCBA after wave soldering and generate thermal inertia anisotropy maps characterizing the anisotropy of solder thermal conduction based on their infrared thermal imaging. S2. Based on the thermal inertia anisotropy map and the geometric morphology of the short-circuit solder joints, target short-circuit solder joints whose thermal inertia anisotropy in the bridging region meets the preset threshold condition are selected. S3. Identify the direction of the minimum thermal inertia principal axis of the bridging region of the target short-circuit weld point and determine it as the direction of the thermodynamic weak interface, and plan a repair strategy that includes the weld nozzle motion trajectory and directional Lorentz force parameters. S4. Control the welding nozzle to move along the welding nozzle trajectory and heat the bridging area to form locally molten solder. At the same time, apply an electromagnetic field corresponding to the directional Lorentz force parameter to drive the molten solder to migrate directionally along the thermodynamic weak interface direction and break the short-circuit bridge. S5. Real-time monitoring of physical quantities characterizing the bridging disconnection. When the bridging is detected to be disconnected, control the welding nozzle to perform negative pressure back suction and simultaneously terminate the application of the electromagnetic field.

2. The PCBA short-circuit solder joint selective repair method based on selective wave soldering according to claim 1, characterized in that, Specifically, S1 is: Images of short-circuit solder joints in PCBA after wave soldering are simultaneously acquired using a visible light camera and an infrared thermal imager to obtain visible light images and time-series multi-frame infrared thermal images of the short-circuit solder joints. Based on the visible light image, the short-circuit bridging region is identified and segmented, and the segmented short-circuit bridging region is spatially mapped onto the multi-frame infrared thermal image to obtain the infrared thermal image sequence of the bridging region. The thermal response curves of temperature change over time for each pixel are extracted from the infrared thermal image sequence. Based on the slope and attenuation characteristics of the temperature rise segment of the thermal response curves, the thermal inertia scalar values ​​of each pixel are calculated, and a thermal inertia scalar distribution map of the bridging region is generated. In the thermal inertia scalar distribution map, the rate of change of thermal inertia is calculated for each pixel along different directions to obtain the thermal inertia direction gradient of each pixel, and the thermal inertia extreme value direction of each pixel is determined according to the thermal inertia direction gradient to construct the thermal inertia vector field of the bridging region. Spatial clustering is performed on the thermal inertia vector field, and the consistent principal axis of the thermal inertia extreme value direction within the clustering region is extracted. The thermal inertia principal axis direction and the thermal inertia anisotropy characterizing the degree of directional difference at each location in the bridging region are then fitted. The thermal inertia principal axis direction and thermal inertia anisotropy are mapped to the geometric space of the bridging region to generate a thermal inertia anisotropy map.

3. The PCBA short-circuit solder joint selective repair method based on selective wave soldering according to claim 1, characterized in that, Specifically, S2 is: The thermal inertia anisotropy values ​​of each pixel in the bridging region are extracted from the thermal inertia anisotropy map. At the same time, the geometric features of the bridging region are extracted from the short-circuit solder joint image to obtain the width distribution of the bridging and the spacing between the solder pads at both ends of the bridging. Based on the width distribution of the bridge, identify the geometric contraction of the bridge extending longitudinally, and delineate the thermally induced separation potential zone at the geometric contraction. Within the thermally induced separation potential region, the regional average value of the thermal inertia anisotropy is calculated and compared with a preset lower limit threshold for anisotropy. When the average value of the region reaches or exceeds the preset lower limit threshold of anisotropy, the corresponding short-circuit solder joint is marked as a candidate repairable solder joint. For the candidate repairable solder joints, the directional deviation angle between the principal axis of thermal inertia in the thermally induced separation potential zone and the longitudinal direction of the bridging is obtained. When the directional deviation angle is less than the preset allowable deviation angle threshold, the corresponding candidate repairable solder joint is determined as the target short-circuit solder joint suitable for selective soldering repair.

4. The PCBA short-circuit solder joint selective repair method based on selective wave soldering according to claim 1, characterized in that, Specifically, S3 is: From the thermal inertia anisotropy map of the target short-circuit weld point, the thermal inertia principal axis direction field and the corresponding anisotropy scalar field of the bridging region are extracted, and the anisotropy scalar field is spatially differentiated to construct an anisotropy gradient vector field that characterizes the degree and direction of anisotropy change. In the anisotropy gradient vector field, starting from the geometric contraction of the bridging region, streamline tracing is performed along the negative anisotropy gradient direction to generate a thermodynamic weak interface path extending from the geometric contraction to the pad safety boundary, such that the tangent direction at each point of the thermodynamic weak interface path corresponds to the principal axis direction of the local minimum thermal inertia. Extract the principal axis directions of thermal inertia at each point along the thermodynamic weak interface path, and select the principal axis direction of thermal inertia with the smallest angle to the tangent direction of the thermodynamic weak interface path as the local weak interface orientation vector at the corresponding point. Determine the ordered set of all local weak interface orientation vectors on the path as the thermodynamic weak interface direction. The thermodynamic weak interface path is spline smoothed, and an offset path is obtained by translating along the normal of the thermodynamic weak interface path according to the preset offset distance between the welding nozzle and the bridging region. The offset path is used as the welding nozzle motion trajectory. At each point along the thermodynamic weak interface path, based on the anisotropy value of each point and the curvature corresponding to the bias path, the current amplitude parameters and magnetic field direction parameters of the directional Lorentz force that match the local weak interface orientation vector of the corresponding point are determined, and synchronously correlated along the welding nozzle movement trajectory according to position to form the directional Lorentz force parameters.

5. The PCBA short-circuit solder joint selective repair method based on selective wave soldering according to claim 4, characterized in that, Based on the anisotropy values ​​at each point and the curvature corresponding to the bias path, the current amplitude parameters and magnetic field direction parameters of the directional Lorentz force that match the local weak interface orientation vector at the corresponding point are determined as follows: Extract each sampling point along the thermodynamic weak interface path at a preset interval, and obtain the local weak interface orientation vector, anisotropy value, radius of curvature at the corresponding position on the bias path corresponding to the sampling point, and preset welding nozzle travel speed at each sampling point. The local weak interface orientation vector at each sampling point is determined as the direction of the target driving force of the molten solder. A local flow coordinate system that varies with the sampling point is established with the direction of the target driving force of the molten solder as the vertical axis and the direction perpendicular to the direction of the target driving force of the molten solder in the pad plane as the horizontal axis. Within the local flow coordinate system, based on the anisotropy value at the sampling point, the thermodynamic migration resistance coefficient at the sampling point is obtained through a preset mapping relationship between anisotropy and thermodynamic migration resistance coefficient; and based on the radius of curvature and nozzle travel speed at the corresponding position on the offset path, the centrifugal induced offset force coefficient generated at the sampling point due to the bending of the nozzle's motion trajectory is calculated. The thermodynamic migration resistance coefficient and the centrifugal induced offset force coefficient are vector summed to obtain the total resistance vector of solder migration at the sampling point. Based on the conductivity parameter of the solder in the molten state and the preset nominal value of the applied magnetic induction intensity, the amplitude parameter of the induced current required to generate the Lorentz force along the direction of the target driving force of the molten solder and with an amplitude equal to the amplitude of the total resistance vector of solder migration is determined. Based on the chirality rule between the direction of the Lorentz force, the direction of the induced current and the direction of the magnetic field, the magnetic field direction parameters that match the amplitude parameters of the induced current at the corresponding sampling point and the direction of the target driving force of the molten solder are determined in the local flow coordinate system, so that the generated directional Lorentz force coincides with the direction of the target driving force of the molten solder. The induced current amplitude parameters and magnetic field direction parameters of each sampling point are matched one-to-one with the position coordinates of the bias path to form directional Lorentz force parameters that are synchronously associated with the welding nozzle movement trajectory.

6. The PCBA short-circuit solder joint selective repair method based on selective wave soldering according to claim 1, characterized in that, Specifically, S4 is: Control the welding nozzle to move to the starting point of the welding nozzle movement trajectory, and preheat the surface of the bridging area with the preset preheating power so that the solder in the bridging area reaches the solid plastic temperature range and forms a solder softening zone. The welding nozzle moves from the starting point to the ending point along the welding nozzle movement trajectory. During the movement, the heating power of the welding nozzle is switched from preheating power to melting power. Local melting zones are formed point by point along the welding nozzle movement path within the solder softening zone. The leading edge of the molten pool in the local melting zone is aligned with the direction of the thermodynamic weak interface. At the synchronous moment of the welding nozzle's movement, the corresponding induced current amplitude parameter and magnetic field direction parameter are retrieved from the directional Lorentz force parameter according to the current position of the welding nozzle, and a transient excitation current corresponding to the induced current amplitude parameter is applied to the electrode pair set on the side of the bridging region. At the same time, the excitation coil built into the welding nozzle is excited to generate a directional transient magnetic field corresponding to the magnetic field direction parameter. The transient excitation current is conducted through the molten solder in the local melting zone to form an induced current. The induced current interacts with the directional transient magnetic field to generate a directional Lorentz force, which drives the molten solder in the local melting zone to migrate along the direction of the thermodynamic weak interface to the outside of the solder nozzle movement trajectory, forming a molten solder migration flow. As the welding nozzle continues to travel, the molten solder migration flow continuously delivers molten solder to the outside of the bridge along the direction of the thermodynamic weak interface, and the width of the bridge cross section continues to narrow until the bridge is completely broken at the geometric contraction part, forming the bridge break end face.

7. The PCBA short-circuit solder joint selective repair method based on selective wave soldering according to claim 1, characterized in that, Specifically, S5 is: As the welding nozzle travels along its trajectory and applies a directional Lorentz force to the bridging region, a high-frequency, low-amplitude probe current signal superimposed on the transient excitation current is continuously applied to the electrode pair, and the response voltage signals at both ends of the electrode pair are acquired simultaneously. The AC impedance component with the same frequency as the high-frequency low-amplitude detection current signal is extracted from the response voltage signal. Based on the real-time change of the AC impedance component in the welding nozzle travel sequence, a dynamic impedance spectrum characterizing the connectivity state of the molten solder in the local melting zone is constructed. The impedance rise trend corresponding to the narrowing process of the bridge cross section in the dynamic impedance spectrum is monitored. When an instantaneous impedance step change is detected in the impedance rise trend and the amplitude of the instantaneous impedance step change exceeds the preset circuit breaker threshold, a bridge disconnection confirmation signal is generated. In response to the bridge disconnection confirmation signal, the transient excitation current and high-frequency low-amplitude detection current signal applied to the electrode pair are immediately turned off, and the negative pressure suction channel at the end of the welding nozzle is activated to form a directional suction flow field at the bridge disconnection end face, which sucks the free molten solder that has migrated to the vicinity of the welding nozzle into the inner cavity of the welding nozzle. A reverse decaying current is applied to the excitation coil built into the welding nozzle according to a preset demagnetization timing curve, so that the magnetic field strength generated by the excitation coil gradually decays from the current directional transient magnetic field level to zero within a predetermined demagnetization time, thus completing the termination of the electromagnetic field application.

8. A PCBA short-circuit solder joint selective repair system based on selective wave soldering, characterized in that, The system includes a memory and a processor. The memory stores a program for selective wave soldering repair of PCBA short-circuit solder joints. When the program for selective wave soldering repair of PCBA short-circuit solder joints is executed by the processor, the steps of the method as described in any one of claims 1 to 7 are implemented.

Images