SMT mounting sequence control method and system

Abstract

The present application relates to the technical field of SMT mounting sequence control, and particularly relates to an SMT mounting sequence control method and system, which comprises the following steps: for each candidate component to be mounted, calculating the moving time cost, physical interference risk cost and solder paste adhesion decay risk cost of the candidate component to be mounted; comprehensively calculating the moving time cost, physical interference risk cost and solder paste adhesion decay risk cost of the candidate component to be mounted to obtain the total cost of each candidate component to be mounted; based on the total cost, iteratively selecting the candidate component to be mounted with the lowest total cost as the next mounting target until all candidate components to be mounted are mounted, thereby generating a globally optimal mounting sequence; and generating mounting instructions executable by a chip mounter according to the globally optimal mounting sequence. The above can effectively balance production efficiency and mounting quality.

Description

Technical Field

[0001] This invention relates to the technical field of SMT placement sequence control, and specifically to an SMT placement sequence control method and system. Background Technology

[0002] In the electronics manufacturing industry, surface mount technology (SMT) is a core process, and its efficiency and quality directly determine the output and product reliability of the entire production line. Among these components, the pick-and-place machine is a key piece of equipment in the SMT production line, responsible for precisely placing electronic components onto designated pads on the printed circuit board (PCB). To maximize production efficiency, the pick-and-place machine's control system typically performs a placement sequence optimization process. The core objective of this process is to plan the shortest path for the placement head, enabling it to pick up and place all components on a PCB as quickly as possible, thereby shortening the production cycle of a single board and increasing throughput per unit time. The control system reads the PCB design information, obtains the coordinates of each component, and combines this with the feeder's position information to calculate a theoretically optimal placement sequence through path planning, then drives the placement head to execute operations in this order.

[0003] However, in actual high-density assembly production, simply pursuing the fastest placement speed can lead to a series of potential quality problems. These problems become particularly prominent when the component layout on the PCB is very compact. Summary of the Invention

[0004] The purpose of this invention is to address the aforementioned shortcomings by proposing an SMT placement sequence control method and system.

[0005] The present invention adopts the following technical solution: A method for controlling the SMT placement sequence, the method comprising the following steps: For each candidate component to be placed, calculate the moving time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate component to be placed. The total cost of each candidate component to be placed is obtained by combining the moving time cost, physical interference risk cost, and solder paste adhesion decay risk cost. Based on the total cost, the candidate component to be placed with the lowest total cost is iteratively selected as the next placement target until all candidate components to be placed are placed, generating the globally optimal placement order; Based on the globally optimal placement order, placement instructions that can be executed by the pick-and-place machine are generated.

[0006] This technical solution comprehensively considers factors such as movement time, physical interference, and solder paste viscosity decay to generate the globally optimal placement sequence, effectively balancing production efficiency and placement quality, and solving the limitations of the single optimization objective in traditional methods.

[0007] Furthermore, the steps for calculating the cost of physical interference risk include the following: During the movement of the placement head, local temperature and humidity data of the area where the placement operation is about to be performed or the area where small components have been placed are acquired; Based on local temperature and humidity data, the predicted viscosity of solder paste on pads where small components have been placed is adjusted. When assessing the physical interference risk cost of placing large components on surrounding small components, the actual holding force of the small components is calculated based on the revised predicted solder paste viscosity, the contact area of ​​the small components, and the weight of the small components. Calculate the holding force vulnerability coefficient based on the actual holding force of the small components that have been placed. By incorporating the holding force vulnerability coefficient into the calculation of physical disturbance risk cost, the corrected physical disturbance risk cost is obtained. When the revised physical interference risk cost exceeds the preset risk threshold, adjust the pressing speed, pressing height, and holding time of large components.

[0008] Furthermore, the steps to synthesize the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate components to be placed, and to obtain the total cost of each candidate component to be placed, include: Set production targets; Adjust the weighting factor of the movement time cost of candidate components to be mounted according to production targets; Adjust the weighting coefficient of physical interference risk costs according to production targets; Adjust the weighting coefficient of solder paste viscosity degradation risk cost according to production targets; The adjusted weighting coefficients for the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost of the candidate components to be mounted are applied to the comprehensive calculation of the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost of the candidate components to be mounted, so as to obtain the total cost of each candidate component to be mounted.

[0009] Furthermore, the steps for calculating the movement time cost of candidate components to be mounted include: Temperature sensors are deployed in key areas of the pick-and-place machine's motion system to collect local temperature data; Collect drive signals and position feedback data during the movement of the placement head; Based on the drive signal and position feedback data, calculate the instantaneous motion damping of the mounting head on the micro-segment of the path; Establish the relationship between local temperature and instantaneous motion damping; Based on the relationship between local temperature and instantaneous motion damping, predict the motion damping on the micro-segment of the path; Calculate the actual time for the mounting head to pass through the micro-segment of the path based on the predicted motion damping; The actual time taken for the placement head to traverse a micro-segment of the path is summed to obtain the movement time cost of the candidate components to be placed.

[0010] Furthermore, the steps of deploying temperature sensors in key areas of the pick-and-place machine's motion system to collect local temperature data include: The deployment location of the temperature sensor is determined based on the sensitivity of the moving parts to the effects of thermal expansion and contraction. The number of temperature sensors to be deployed is determined based on the placement head's movement frequency and dwell time. Based on the determined number of temperature sensors to be deployed, multiple temperature sensors are deployed in key areas of the linear guide rail, ball screw, and drive motor of the XY motion system of the pick-and-place machine to collect local temperature data. The deployment location of each temperature sensor is mapped to the area of ​​the motion path it monitors, so as to establish a correspondence between the motion path area and the local temperature data collected by the temperature sensor. When the mounting head moves to a specific path segment, it calls the data collected by the temperature sensor associated with that path segment according to the established correspondence, evaluates the local temperature gradient of that path segment, and uses it to predict the motion damping on the micro-segment of the path.

[0011] Furthermore, the steps to establish the relationship between local temperature and instantaneous motion damping include: Temperature sensors, humidity sensors, and lubricant status sensors are deployed in key areas of the pick-and-place machine's motion system to collect local temperature, humidity, and lubricant status data. Collect drive signals and position feedback data during the movement of the placement head; Based on the drive signal and position feedback data, calculate the instantaneous motion damping of the mounting head on the micro-segment of the path; The local temperature, humidity, and lubricant status data are matched with instantaneous motion damping using timestamps and spatial location correlation. Based on the correlation results, the nonlinear influence patterns of local temperature, humidity, and lubricant status on instantaneous motion damping are identified to improve the accuracy of motion damping on subsequent predicted path micro-segments. Based on the nonlinear influence mode, the relationship parameters between local temperature and instantaneous motion damping are dynamically adjusted.

[0012] Furthermore, based on the association results, the method also includes the following steps: Based on the correlation results, the dynamic evolution trend of the nonlinear influence mode of equipment aging, production batch differences or maintenance cycle on local temperature, humidity, lubricant status and instantaneous motion damping is analyzed. Based on the dynamic evolution trend, the parameters of the nonlinear influence mode are periodically updated to adapt to the complex relationship that evolves over time, thereby improving the accuracy of motion damping on subsequent predicted path micro-segments.

[0013] Furthermore, the steps of periodically updating the parameters of the nonlinear influence mode based on the dynamic evolution trend include: During the operation of the pick-and-place machine, continuously monitor the equipment aging status, production batch information, and maintenance cycle data; Based on equipment aging status, production batch information, and maintenance cycle data, assess the dynamic evolution rate of parameters of nonlinear influence modes. When the dynamic evolution rate exceeds the preset rate threshold, the parameters of the nonlinear influence mode are updated, and the update magnitude of the parameters of the nonlinear influence mode is adjusted according to the dynamic evolution rate. When the dynamic evolution rate is lower than the preset rate threshold, the update cycle of the parameters of the nonlinear influence mode is extended and the update amplitude of the parameters of the nonlinear influence mode is reduced. When updating the parameters of the nonlinear influence model, a weighted average method is used to fuse the parameters of the historical nonlinear influence model and the parameters of the latest calculated nonlinear influence model. The weighted average method dynamically adjusts the weights based on the dynamic evolution rate and data confidence.

[0014] Furthermore, the method also includes: During the operation of the pick-and-place machine, the generation time of parameters of historical nonlinear influence modes is continuously monitored; Based on the generation time of the parameters of the historical nonlinear influence model, calculate the time-dependent decay coefficient of the parameters of the historical nonlinear influence model. The weights of the parameters of historical nonlinear influence modes and the weights of the parameters of the latest calculated nonlinear influence modes are dynamically adjusted based on the dynamic evolution rate, data confidence, and timeliness decay coefficient. The parameters of the historical nonlinear influence model are multiplied by their corresponding weights, and the parameters of the latest calculated nonlinear influence model are multiplied by their corresponding weights. The two products are then added together to obtain the parameters of the fused nonlinear influence model.

[0015] Through this technical solution, this application can more finely adjust the weights of historical and latest parameters by introducing a timeliness decay coefficient and combining dynamic evolution rate and data confidence, making the fused nonlinear influence model parameters more timely and accurate, and further improving prediction accuracy.

[0016] This application also discloses an SMT placement sequence control system, applied to an SMT placement sequence control method, the system comprising: The calculation module calculates the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost for each candidate component to be placed. The synthesis module combines the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate components to be placed, and obtains the total cost of each candidate component to be placed. The sequence generation module iteratively selects the candidate component to be placed with the lowest total cost as the next placement target based on the total cost, until all candidate components to be placed are placed, generating the globally optimal placement sequence. The instruction generation module generates placement instructions that can be executed by the pick-and-place machine based on the globally optimal placement order.

[0017] Through this technical solution, this application can effectively implement the SMT placement sequence control method through modular design, providing a complete and operable solution for actual production, thereby improving the intelligence level and overall efficiency of the production line.

[0018] This application balances production efficiency and placement quality, avoiding path redundancy caused by traditional "large-to-small" rules, while also mitigating solder paste viscosity degradation due to extended placement time. Therefore, this application significantly improves the overall efficiency and product yield of SMT production lines, providing a more intelligent and reliable placement sequence optimization solution for high-density electronic assembly.

[0019] To further understand the features and technical content of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are for reference and illustration only and are not intended to limit the present invention. Attached Figure Description

[0020] Figure 1 This is a flowchart of an SMT placement sequence control method according to the present invention; Figure 2 This is a schematic diagram of the structure of an SMT placement sequence control system according to the present invention. Detailed Implementation

[0021] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can understand the advantages and effects of the present invention from the content disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the spirit of the present invention. Furthermore, the accompanying drawings of the present invention are for simple illustrative purposes only and are not depictions of actual dimensions; this is stated in advance. The following embodiments will further describe the relevant technical content of the present invention in detail, but the disclosed content is not intended to limit the scope of protection of the present invention.

[0022] This embodiment provides an SMT placement sequence control method and system, combined with Figure 1 and Figure 2 As shown.

[0023] refer to Figure 1 A method for controlling the SMT placement sequence, the method comprising the following steps: For each candidate component to be placed, calculate the moving time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate component to be placed. The total cost of each candidate component to be placed is obtained by combining the moving time cost, physical interference risk cost, and solder paste adhesion decay risk cost. Based on the total cost, the candidate component to be placed with the lowest total cost is iteratively selected as the next placement target until all candidate components to be placed are placed, generating the globally optimal placement order; Based on the globally optimal placement order, placement instructions that can be executed by the pick-and-place machine are generated.

[0024] The "SMT placement sequence control method" mentioned in this application refers to a series of algorithms and operational steps used to optimize the component placement sequence on a surface mount technology (SMT) production line. This method aims to balance production efficiency and product quality by generating an optimal component placement sequence through a comprehensive evaluation of various cost factors, guiding the pick-and-place machine to perform precise, efficient, and non-destructive placement operations. "Candidate components to be placed" refers to all components that have not yet been placed onto the printed circuit board (PCB) in the current placement task. These components are the objects that the system needs to evaluate and select when generating the placement sequence. "Movement time cost" refers to the time required for the pick-and-place machine's placement head to move from its current position to the target component pick-up position and then to the target component placement position, typically related to factors such as path length, movement speed, and acceleration / deceleration performance. "Physical interference risk cost" refers to the risk of physical impact or positional displacement to already placed or unplaced components during the placement process due to the movement of the placement head and component placement, especially significant on PCBs with high-density, mixed-size component layouts. "Solder paste viscosity degradation risk cost" refers to the risk that the viscosity of solder paste on the pads will gradually decrease with time, temperature, humidity, and other environmental factors when exposed to air, thus affecting the component's holding power. This risk may cause components to shift or fall off during subsequent handling. "Total cost" is a single evaluation metric obtained by weighting or comprehensively calculating the above three costs, used to quantify the overall advantages and disadvantages of each candidate component to be placed. "Globally optimal placement sequence" refers to finding the sequence that minimizes the total cost among all possible placement sequences through an iterative optimization process, maximizing the balance between production efficiency, physical interference risk, and solder paste viscosity degradation risk. "Pick-and-place machine executable placement instructions" refers to translating the generated globally optimal placement sequence into machine language or G-code that the pick-and-place machine control system can recognize and execute, driving the pick-and-place machine to complete the placement operation in a predetermined order.

[0025] When implementing the SMT placement sequence control method of this application, it is first necessary to calculate the cost of each candidate component to be placed. For example, the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate components to be placed can be calculated in the following ways.

[0026] To calculate the movement time cost, a motion model of the pick-and-place machine can be pre-established. This model can estimate the theoretical movement time based on parameters such as the path length of the placement head from its current position to the target component pickup position and then to the placement position, the maximum speed of the pick-and-place machine, acceleration, and deceleration. For example, by dividing the movement path of the placement head into multiple micro-segments and summing the movement time of each micro-segment, the total movement time cost can be obtained.

[0027] To calculate the cost of physical interference risk, factors such as component size, weight, placement head pressing speed, and the holding force of surrounding mounted components can be considered. For example, when a large component is about to be placed, the impact of potential airflow impact or mechanical vibration during its pressing process on surrounding small mounted components can be assessed. This impact can be quantified using an empirical formula or simulation model; for instance, by taking the component's size, weight, pressing speed, and distance from surrounding components as inputs, a physical interference risk index can be output.

[0028] To calculate the risk cost of solder paste viscosity degradation, a solder paste viscosity degradation model can be established. This model can predict the degree of viscosity degradation based on factors such as solder paste exposure time, ambient temperature, and humidity. For example, experiments can be conducted beforehand to obtain the viscosity change curves of solder paste under different exposure times, temperatures, and humidity levels, and then these data can be fitted into a mathematical model. In actual calculations, based on the exposure time of the solder paste on the pads corresponding to the current components and environmental parameters, the degree of viscosity degradation can be predicted and converted into a risk cost.

[0029] After calculating the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost for each candidate component to be placed, these costs need to be aggregated to obtain the total cost for each candidate component. One approach is to use a weighted summation method. For example, a weighting coefficient can be assigned to each of the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost. Then, each cost is multiplied by its corresponding weighting coefficient, and finally, all weighted costs are summed to obtain the total cost. These weighting coefficients can be adjusted according to actual production needs. For example, if the current production task has high efficiency requirements, the weight of the movement time cost can be appropriately increased; if quality requirements are high, the weights of the physical interference risk cost and the solder paste viscosity degradation risk cost can be increased.

[0030] After obtaining the total cost of each candidate component to be placed, this application iteratively selects the candidate component with the lowest total cost as the next placement target based on the total cost, until all candidate components are placed, generating a globally optimal placement order. For example, a greedy algorithm or a heuristic algorithm can be used for iterative selection. In each iteration, the system evaluates all unplaced components, calculates their total cost, and then selects the component with the lowest total cost as the placement target for the current step. This process is repeated until all components are selected and arranged in the placement order. For example, in the first step, the system calculates the total cost of all components to be placed and selects the lowest cost component A; in the second step, the system recalculates the total cost of the remaining components to be placed (considering that component A has been placed, which may affect the cost of subsequent components), selects the lowest cost component B, and so on, ultimately forming a complete placement order.

[0031] Finally, based on the globally optimal placement sequence, placement instructions executable by the pick-and-place machine are generated. For example, the generated placement sequence can be converted into G-code or a specific machine instruction format that the pick-and-place machine control system can recognize. These instructions will contain detailed parameters such as the pick-up position, placement position, placement speed, pressure drop height, and holding pressure time for each component. For example, an instruction sequence might include "pick up component X at feeder Y position", "move to PCB coordinates (x, y)", "press down at speed V", and "hold pressure for T seconds". After these instructions are transmitted to the pick-and-place machine control system, the machine can be driven to perform automated placement operations according to the globally optimal placement sequence.

[0032] This application further proposes steps for calculating the cost of physical interference risks, including: During the movement of the placement head, local temperature and humidity data of the area where the placement operation is about to be performed or the area where small components have been placed are acquired; Based on local temperature and humidity data, the predicted viscosity of solder paste on pads where small components have been placed is adjusted. When assessing the physical interference risk cost of placing large components on surrounding small components, the actual holding force of the small components is calculated based on the revised predicted solder paste viscosity, the contact area of ​​the small components, and the weight of the small components. Calculate the holding force vulnerability coefficient based on the actual holding force of the small components that have been placed. By incorporating the holding force vulnerability coefficient into the calculation of physical disturbance risk cost, the corrected physical disturbance risk cost is obtained. When the revised physical interference risk cost exceeds the preset risk threshold, adjust the pressing speed, pressing height, and holding time of large components.

[0033] Specifically, during the placement head's movement, miniature temperature and humidity sensors can be deployed near the placement head or in specific areas of the PCB board to collect real-time local temperature and humidity data for the area about to be placed or where small components have already been placed. This data reflects the actual environmental conditions of the solder paste. Correcting the predicted solder paste viscosity on the pads of placed small components based on local temperature and humidity data involves dynamically adjusting the initially set predicted solder paste viscosity using a pre-established model of the relationship between solder paste viscosity and temperature and humidity. For example, when local humidity or temperature is high, the solder paste viscosity may decrease, and the predicted value will be corrected accordingly. In practical applications, when assessing the physical interference risk and cost of placing large components on surrounding small components, the actual holding force of the small components is calculated based on the corrected predicted solder paste viscosity, the contact area of ​​the placed small components, and the weight of the placed small components. Actual holding force can be understood as the actual adhesion strength of solder paste to components under current temperature and humidity conditions. Its calculation comprehensively considers solder paste viscosity, the contact area between the component and the pad, and the component's own weight. Furthermore, based on the actual holding force of the placed small components, a holding force vulnerability coefficient is calculated. The holding force vulnerability coefficient is a quantitative indicator used to represent the probability that placed components will shift or detach when subjected to external impact or pressure. The lower the holding force, the higher the vulnerability coefficient. Therefore, the holding force vulnerability coefficient is introduced into the calculation of physical interference risk cost, resulting in a corrected physical interference risk cost. This means that the calculation of physical interference risk cost is no longer based solely on static factors such as component size and spacing, but dynamically considers the actual holding state of the placed components. When the corrected physical interference risk cost exceeds a preset risk threshold, the system will automatically adjust the pressing speed, pressing height, and holding time of large components. For example, the pressing speed can be appropriately reduced, the pressing height decreased, or the holding time extended to reduce the impact on surrounding components and ensure the stability and safety of the mounting process.

[0034] In some preferred embodiments, it is assumed that during the SMT placement process, the placement head is preparing to place a large component (e.g., a BGA chip), while several small components (e.g., 0402 resistors) are already placed around it. At this time, a miniature temperature and humidity sensor deployed near the placement area collects real-time data showing a local ambient temperature of 30°C and a relative humidity of 70%. Based on a preset model of the relationship between solder paste viscosity and temperature and humidity, the system lowers the previously set predicted solder paste viscosity value under standard conditions (25°C, 50%RH) to reflect the potential decrease in solder paste viscosity under higher temperature and humidity. Based on the corrected predicted solder paste viscosity value, the contact area of ​​the 0402 resistors, and their own weight, the system calculates the actual holding force of these small components. For example, if the calculated actual holding force is lower than a certain threshold, a higher holding force vulnerability coefficient is calculated. Subsequently, this vulnerability coefficient is incorporated into the calculation of the physical interference risk cost of placing the large component, resulting in a corrected physical interference risk cost. If the revised risk cost exceeds a preset risk threshold (e.g., indicating a high probability of displacement of small components), the system will immediately send a command to the pick-and-place machine controller to reduce the BGA chip pressing speed from the usual 50mm / s to 30mm / s, adjust the pressing height from 0.1mm to 0.05mm, and extend the holding time from 0.5 seconds to 1 second. These dynamic adjustments significantly reduce the impact of airflow and mechanical vibration generated during BGA chip pressing on surrounding fragile small components, effectively preventing component displacement or detachment and ensuring the accuracy and reliability of the placement process.

[0035] The steps to synthesize the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost of candidate components to obtain the total cost of each candidate component to be placed include: Set production targets; Adjust the weighting factor of the movement time cost of candidate components to be mounted according to production targets; Adjust the weighting coefficient of physical interference risk costs according to production targets; Adjust the weighting coefficient of solder paste viscosity degradation risk cost according to production targets; The adjusted weighting coefficients for the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost of the candidate components to be mounted are applied to the comprehensive calculation of the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost of the candidate components to be mounted, so as to obtain the total cost of each candidate component to be mounted.

[0036] Setting production goals can be understood as clarifying the production focus for the current batch or stage based on actual production needs. For example, production goals may include "maximizing production efficiency," "minimizing the risk of physical interference," "optimizing solder paste adhesion retention," or "balancing efficiency and quality." These goals can be manually entered by operators or automatically obtained through the production management system.

[0037] Furthermore, adjusting the weighting coefficients of various costs based on production objectives refers to dynamically assigning different levels of importance to movement time costs, physical interference risk costs, and solder paste viscosity decay risk costs according to the set production goals. For example, when the production goal is "maximizing production efficiency," the weighting coefficient of movement time costs can be increased, while the weighting coefficients of physical interference risk costs and solder paste viscosity decay risk costs can be decreased; when the production goal is "minimizing physical interference risk," the weighting coefficient of physical interference risk costs can be increased. These adjustments to weighting coefficients can be achieved through a pre-set rule base, machine learning model, or expert system.

[0038] Therefore, applying the adjusted weighting coefficients of various costs to the overall calculation means that when calculating the total cost of each candidate component to be mounted, a fixed weight is no longer used, but a dynamic weight that matches the current production target is adopted. The formula for calculating the total cost can be expressed as: Total Cost = W_Time * Mobility Time Cost + W_Physical Cost * Physical Interference Risk Cost + W_Solder Paste * Solder Paste Stickiness Decay Risk Cost, where W_Time, W_Physical Cost, and W_Solder Paste are the weighting coefficients for the adjusted mobility time cost, physical interference risk cost, and solder paste stickiness decay risk cost, respectively. The purpose is to ensure that the final mounting sequence better serves the current production target.

[0039] This application's solution sets production targets and dynamically adjusts the weighting coefficients of the movement time cost, physical interference risk cost, and solder paste viscosity decay risk cost of candidate components to be mounted based on these targets. This allows the total cost calculation to flexibly reflect the current production priorities. Specifically, once a production target is set, the system intelligently adjusts the proportion of each cost in the total cost according to that target. For example, if the current production task has extremely high efficiency requirements, the weighting coefficient of movement time cost will be significantly increased, making the movement time cost have a greater impact on the final result in the total cost calculation, thus guiding the system to prioritize the mounting path with the shortest movement time. Conversely, if the current task has higher requirements for quality and reliability, the weighting coefficients of physical interference risk cost and solder paste viscosity decay risk cost will be increased, prompting the system to prioritize avoiding risks that may lead to component damage or solder paste failure when generating the mounting sequence. This dynamic adjustment mechanism ensures that the total cost calculation is no longer static but can adapt to different production scenarios and needs, thus providing a more guiding basis for subsequent iterations to select the candidate components to be mounted with the lowest total cost.

[0040] In some preferred embodiments, suppose an SMT production line needs to mount a batch of circuit boards containing various components.

[0041] Scenario 1: The production goal is set as "maximizing production efficiency." In this scenario, the system will adjust the weighting coefficients for the movement time cost of candidate components to 0.7, the physical interference risk cost to 0.2, and the solder paste adhesion decay risk cost to 0.1. When calculating the total cost of each candidate component, movement time cost will dominate, prompting the system to prioritize components that significantly shorten the placement head's movement path as the next placement target, thereby minimizing the overall placement time.

[0042] Scenario 2: The production objective is set as "minimizing physical interference risk." In this case, the system will adjust the weighting coefficients for physical interference risk cost to 0.6, movement time cost to 0.2, and solder paste adhesion decay risk cost to 0.2. When evaluating total cost, the system will pay more attention to potential physical interference caused by components during placement, such as avoiding high-voltage placement of large components near already placed small components, thereby effectively reducing the risk of component damage or displacement.

[0043] Scenario 3: The production goal is set as "optimizing solder paste viscosity retention". In this scenario, the system will adjust the weighting coefficient for solder paste viscosity degradation risk cost to 0.5, the weighting coefficient for movement time cost to 0.3, and the weighting coefficient for physical interference risk cost to 0.2. Under this scenario, the system will prioritize placement paths and sequences with lower solder paste viscosity degradation risk, such as prioritizing the placement of components with high solder paste viscosity requirements or those susceptible to environmental influences, to ensure the activity and viscosity of the solder paste before placement, thereby improving soldering quality.

[0044] As can be seen from the above examples, the solution proposed in this application can flexibly adjust the weight of various costs according to different production goals, making the calculation of total cost more targeted, thereby generating a globally optimal mounting sequence that better meets the current production needs.

[0045] This application further proposes steps for calculating the movement time cost of candidate components to be mounted, including: Temperature sensors are deployed in key areas of the pick-and-place machine's motion system to collect local temperature data; Collect drive signals and position feedback data during the movement of the placement head; Based on the drive signal and position feedback data, calculate the instantaneous motion damping of the mounting head on the micro-segment of the path; Establish the relationship between local temperature and instantaneous motion damping; Based on the relationship between local temperature and instantaneous motion damping, predict the motion damping on the micro-segment of the path; Calculate the actual time for the mounting head to pass through the micro-segment of the path based on the predicted motion damping; The actual time taken for the placement head to traverse a micro-segment of the path is summed to obtain the movement time cost of the candidate components to be placed.

[0046] Specifically, temperature sensors are deployed in key areas of the pick-and-place machine's motion system to monitor environmental factors affecting motion performance in real time. These key areas may include, but are not limited to, the linear guides, ball screws, and drive motors of the XY motion system. Temperature changes in these components directly affect their mechanical properties and frictional characteristics. By collecting local temperature data, important environmental parameters can be provided for subsequent motion damping prediction.

[0047] Simultaneously, collecting drive signals and position feedback data during the placement head's movement is crucial for obtaining detailed information about the head's actual motion. The drive signals reflect the torque or current applied by the motor, while the position feedback data provides precise position and velocity information of the placement head in space. This data forms the basis for calculating motion damping.

[0048] Specifically, based on the drive signal and position feedback data, the instantaneous motion damping of the mounting head on the micro-segment of the path is calculated. Instantaneous motion damping refers to the resistance experienced by the mounting head during movement at a specific moment and on a specific micro-segment of the path. This damping value can be obtained by analyzing the relationship between the driving force and the actual motion response, for example, by comparing the deviation between the theoretical driving force and the actual acceleration or velocity.

[0049] Furthermore, the relationship between local temperature and instantaneous motion damping needs to be established. Since temperature changes cause thermal expansion and contraction of mechanical components and changes in lubricant viscosity, thus affecting motion damping, a model can be established using historical data or experimental calibration to describe the variation of instantaneous motion damping under different local temperature conditions.

[0050] Therefore, based on the relationship between local temperature and instantaneous motion damping, the motion damping on the micro-segment of the path can be predicted. In the actual placement process, by using real-time collected local temperature data and combining it with the established relationship model, the motion damping that the placement head will encounter on the micro-segment of the path it is about to pass through can be accurately predicted.

[0051] Subsequently, based on the predicted motion damping, the actual time required for the placement head to traverse the micro-segment of the path is calculated. Once the motion damping of a specific micro-segment of the path is predicted, the actual time required for the placement head to traverse that micro-segment can be calculated using a dynamic model, taking into account parameters such as the mass and driving force of the placement head.

[0052] Finally, by summing up the actual time taken for the placement head to traverse all relevant path segments, the movement time cost of the candidate component to be placed can be obtained. This accumulated value represents the total time required for the placement head to move from its current position to the target placement position, and this time cost fully considers actual mechanical and environmental factors.

[0053] In some preferred embodiments, it is assumed that multiple components need to be mounted on a PCB board on an SMT production line. When calculating the movement time cost of the next candidate component to be mounted, multiple temperature sensors are first deployed in key areas of the pick-and-place machine's XY motion system, such as the linear guides, ball screws, and drive motors. For example, one temperature sensor is deployed at the start, midpoint, and end point of the X-axis guide, and similar sensors are deployed at similar locations on the Y-axis guide and ball screw to collect local temperature data in these areas in real time. Simultaneously, the pick-and-place machine's controller continuously collects drive signals (such as motor current) and position feedback data (such as encoder readings) along the path of the placement head from its current position to the target placement position.

[0054] For example, when the placement head moves from point A to point B, the path is divided into multiple micro-segments. For each micro-segment, the system calculates the instantaneous motion damping of the placement head in that micro-segment based on the collected drive signals and position feedback data, using a pre-defined dynamic model. Simultaneously, the system accesses local temperature data collected by a temperature sensor associated with the spatial location of that micro-segment. Through a pre-established lookup table or mathematical model describing the relationship between local temperature and instantaneous motion damping, the system can correct or predict the motion damping in that micro-segment based on the current local temperature. For example, if the local temperature of a micro-segment is high, the predicted motion damping may decrease accordingly, and vice versa.

[0055] Once the motion damping of each micro-segment is predicted, the system can more accurately calculate the actual time required for the placement head to traverse that micro-segment. For example, in micro-segments with lower damping, the placement head may traverse at a faster speed; while in micro-segments with higher damping, it may require a longer acceleration or deceleration time. Finally, by summing up the actual passage times of all micro-segments, the movement time cost of the candidate component to be placed is obtained. This cost value serves as a crucial input for calculating the total cost, iteratively selecting the next placement target, thereby generating a more accurate and efficient globally optimal placement sequence. In this way, even under conditions of fluctuating ambient temperature or slight variations in the performance of mechanical components, the accuracy of movement time cost calculations is ensured, thus optimizing overall placement efficiency.

[0056] The steps for deploying temperature sensors in key areas of the pick-and-place machine's motion system and collecting local temperature data include: The deployment location of the temperature sensor is determined based on the sensitivity of the moving parts to the effects of thermal expansion and contraction. The number of temperature sensors to be deployed is determined based on the placement head's movement frequency and dwell time. Based on the determined number of temperature sensors to be deployed, multiple temperature sensors are deployed in key areas of the linear guide rail, ball screw, and drive motor of the XY motion system of the pick-and-place machine to collect local temperature data. The deployment location of each temperature sensor is mapped to the area of ​​the motion path it monitors, so as to establish a correspondence between the motion path area and the local temperature data collected by the temperature sensor. When the mounting head moves to a specific path segment, it calls the data collected by the temperature sensor associated with that path segment according to the established correspondence, evaluates the local temperature gradient of that path segment, and uses it to predict the motion damping on the micro-segment of the path.

[0057] Specifically, the sensitivity of moving parts to thermal expansion and contraction refers to the degree to which the size, shape, or performance of each component in the placement machine's motion system changes under temperature variations. For example, for precision transmission components such as linear guides and ball screws, thermal expansion and contraction have a more significant impact on motion accuracy and damping. Therefore, the placement of temperature sensors should prioritize these highly sensitive areas.

[0058] The number of temperature sensors deployed is determined based on the placement head's movement frequency and dwell time. For example, in areas with high placement head movement frequency or long dwell time, heat accumulation and dissipation may lead to larger local temperature fluctuations, thus requiring more temperature sensors for more precise temperature monitoring. Conversely, in areas with low movement frequency or short dwell time, the number of temperature sensors can be appropriately reduced.

[0059] In practical applications, based on a predetermined number of temperature sensors, multiple temperature sensors are deployed in key areas of the XY motion system of the pick-and-place machine, including the linear guide rails, ball screws, and drive motors, to collect localized temperature data. These key areas are the main factors affecting the damping of the placement head's motion; accurate temperature monitoring of these areas helps to more accurately assess the operating status of the motion system.

[0060] Furthermore, the deployment location of each temperature sensor is mapped to the movement path area it monitors, establishing a correspondence between the movement path area and the local temperature data collected by the temperature sensor. This mapping can be stored in a database or configuration file, and its purpose is to ensure that when the mounting head moves to a specific path segment, the local temperature data corresponding to that path segment can be accurately obtained.

[0061] Therefore, when the placement head moves to a specific path segment, it retrieves data collected by the temperature sensor associated with that path segment based on the established correspondence, and evaluates the local temperature gradient of that path segment. The local temperature gradient refers to the rate of temperature change with spatial location within a certain path segment, reflecting the uniformity of temperature distribution and potential thermal deformation risk within that segment. This local temperature gradient data is used to predict motion damping on micro-segments of the path.

[0062] This application further proposes steps for establishing the relationship between local temperature and instantaneous motion damping, including: Temperature sensors, humidity sensors, and lubricant status sensors are deployed in key areas of the pick-and-place machine's motion system to collect local temperature, humidity, and lubricant status data. Collect drive signals and position feedback data during the movement of the placement head; Based on the drive signal and position feedback data, calculate the instantaneous motion damping of the mounting head on the micro-segment of the path; The local temperature, humidity, and lubricant status data are matched with instantaneous motion damping using timestamps and spatial location correlation. Based on the correlation results, the nonlinear influence patterns of local temperature, humidity, and lubricant status on instantaneous motion damping are identified to improve the accuracy of motion damping on subsequent predicted path micro-segments. Based on the nonlinear influence mode, the relationship parameters between local temperature and instantaneous motion damping are dynamically adjusted.

[0063] Specifically, in key areas of the pick-and-place machine's motion system, in addition to temperature sensors to collect local temperature data, humidity sensors and lubricant condition sensors are also deployed. The humidity sensor monitors the air humidity in these key areas in real time, as humidity changes directly affect lubricant performance and the coefficient of friction of moving parts. The lubricant condition sensor is configured to assess key indicators such as lubricant viscosity, cleanliness, and the presence of wear particles, which are important factors affecting motion damping. These sensors are strategically deployed near key friction and moving parts of the XY motion system, such as linear guides, ball screws, and drive motors, to ensure that the collected data accurately reflects the local environment and mechanical condition.

[0064] The process involves collecting drive signals and position feedback data during the movement of the placement head, and calculating the instantaneous motion damping of the placement head on the micro-segment of the path based on these data. The principle is similar to the above implementation method. By analyzing the motor drive current, voltage, and position and speed data fed back by the encoder, the instantaneous damping force or damping coefficient experienced by the placement head on a specific micro-segment of the path can be accurately deduced.

[0065] Furthermore, local temperature, humidity, and lubricant condition data are time-stamped and spatially correlated with instantaneous motion damping. Time-stamp matching ensures precise alignment between multi-dimensional sensor data collected at the same moment or within a very short time interval and instantaneous motion damping data. Spatial correlation maps the deployment locations of each sensor to specific micro-segments along the mounting head's movement path, ensuring accurate correlation between the instantaneous motion damping calculated on a specific micro-segment and the corresponding local environmental (temperature, humidity) and mechanical (lubricant condition) data. The aim is to provide a reliable and consistent data foundation for subsequent identification of the combined effects of multiple factors on damping.

[0066] Based on the aforementioned correlation results, the nonlinear influence patterns of local temperature, humidity, and lubricant state on instantaneous motion damping are identified. Specifically, advanced machine learning algorithms, such as neural networks, support vector machines, or decision trees, can be used to train multi-dimensional data that has been time-stamped and spatially correlated. These algorithms can learn from complex input data and identify the nonlinear relationships between temperature, humidity, lubricant state, and instantaneous motion damping. For example, within specific temperature and humidity ranges, changes in lubricant viscosity may have a more significant impact on damping. The aim is to more accurately understand and model these complex interactions, thereby improving the accuracy of motion damping prediction.

[0067] In practical applications, the relationship parameters between local temperature and instantaneous motion damping are dynamically adjusted based on the identified nonlinear influence patterns. Once a nonlinear influence pattern is identified, the system can use this pattern to dynamically correct or adjust the parameters in the original relationship model between local temperature and instantaneous motion damping, based on real-time data on local temperature, humidity, and lubricant condition. For example, when humidity is high, parameters related to lubricant performance in the model may be adjusted to reflect the damping change caused by the decrease in lubricant viscosity. The aim is to enable the established relationship model to adapt to the constantly changing operating environment and mechanical state in real time, thereby further improving the accuracy of motion damping prediction.

[0068] In some preferred embodiments, it is assumed that on an SMT production line, prolonged operation of the pick-and-place machine leads to increased local ambient temperature, fluctuating workshop humidity, and slight aging of the lubricant due to prolonged use. In this situation, predicting motion damping solely based on local temperature may result in significant deviations due to neglecting the influence of humidity and lubricant condition. The solution of this application, by deploying humidity and lubricant condition sensors, can collect this data in real time. For example, when humidity increases, the system identifies the nonlinear influence pattern of humidity on lubricant viscosity and dynamically adjusts the relationship parameters between local temperature and instantaneous motion damping accordingly, ensuring that the predicted motion damping accurately reflects the impact of humidity changes. Similarly, when the lubricant condition sensor detects signs of lubricant aging, the system corrects the damping prediction model based on pre-learned nonlinear patterns, thereby ensuring high accuracy in motion damping prediction even under complex changes in environmental and mechanical conditions. Consequently, the actual time for the placement head to traverse micro-segments of the path can be calculated more accurately, thereby optimizing movement time costs and ultimately generating a more reliable globally optimal placement sequence.

[0069] Based on the association results, the method also includes the following steps: Based on the correlation results, the dynamic evolution trend of the nonlinear influence mode of equipment aging, production batch differences or maintenance cycle on local temperature, humidity, lubricant status and instantaneous motion damping is analyzed. Based on the dynamic evolution trend, the parameters of the nonlinear influence mode are periodically updated to adapt to the complex relationship that evolves over time, thereby improving the accuracy of motion damping on subsequent predicted path micro-segments.

[0070] Specifically, the aforementioned "correlation results" refer to the corresponding dataset between local temperature, humidity, lubricant status data and instantaneous motion damping obtained through timestamp matching and spatial location correlation in the above embodiments. "Equipment aging" can be understood as the wear, increased clearance, or performance degradation of components such as linear guides, ball screws, and drive motors in the placement machine's motion system. These signs of aging can be assessed by monitoring equipment operating time, cumulative travel distance, failure rate, or data collected by sensors (e.g., vibration sensors, current sensors). "Production batch variation" refers to minor changes in the physical properties (e.g., dimensional accuracy, weight, solder paste viscosity, flowability) of different batches of components or solder paste. These changes may affect the interaction between the placement head and components, as well as the curing characteristics of the solder paste, thus indirectly affecting motion damping. This can be identified through production batch numbers, material inspection reports, and other data. "Maintenance cycle" refers to the time interval between regular equipment maintenance, lubricant replacement, or component calibration. Maintenance operations alter the frictional characteristics and lubrication status of the motion system, thereby affecting motion damping. This can be tracked through maintenance records or maintenance plans. "Dynamic evolution trend" refers to the regular or irregular changes in the parameters or intrinsic relationships of the aforementioned nonlinear influence patterns as time, operating conditions, or external factors change. This trend can be analyzed using methods such as time series analysis, regression analysis, and machine learning models (e.g., Long Short-Term Memory Network LSTM, Recurrent Neural Network RNN) to identify the drift, decay, or enhancement of patterns through historical data.

[0071] Here, "parameters of the nonlinear influence mode" refers to the coefficients, weights, or thresholds in the mathematical model (e.g., polynomial model, neural network model) established in the above embodiments to describe the relationship between local temperature, humidity, lubricant state, and instantaneous motion damping. "Periodic updates" can be understood as updates occurring at preset time intervals (e.g., daily, weekly, monthly) or event-triggered, such as triggering an update immediately when a dynamic evolution trend exceeds a preset threshold. "Adapting to complex relationships evolving over time" refers to updating parameters to enable the model to better fit the current and future actual operating states, thereby avoiding prediction errors caused by the model's disconnect from reality. "Improving the accuracy of motion damping on subsequent predicted path micro-segments" refers to continuously optimizing model parameters to ensure more accurate predictions of motion damping of the mounting head on different path micro-segments, thus providing a more reliable basis for calculating movement time costs.

[0072] This application further proposes a step for periodically updating the parameters of the nonlinear influence mode based on dynamic evolution trends, including: During the operation of the pick-and-place machine, continuously monitor the equipment aging status, production batch information, and maintenance cycle data; Based on equipment aging status, production batch information, and maintenance cycle data, assess the dynamic evolution rate of parameters of nonlinear influence modes. When the dynamic evolution rate exceeds the preset rate threshold, the parameters of the nonlinear influence mode are updated, and the update magnitude of the parameters of the nonlinear influence mode is adjusted according to the dynamic evolution rate. When the dynamic evolution rate is lower than the preset rate threshold, the update cycle of the parameters of the nonlinear influence mode is extended and the update amplitude of the parameters of the nonlinear influence mode is reduced. When updating the parameters of the nonlinear influence model, a weighted average method is used to fuse the parameters of the historical nonlinear influence model and the parameters of the latest calculated nonlinear influence model. The weighted average method dynamically adjusts the weights based on the dynamic evolution rate and data confidence.

[0073] Specifically, during the operation of the pick-and-place machine, continuous monitoring of equipment aging status, production batch information, and maintenance cycle data aims to identify key factors influencing the evolution of parameters affecting nonlinear influence modes. Equipment aging status reflects long-term changes such as wear and tear of mechanical components and degradation of electrical performance; production batch information reveals potential initial differences between different batches of components or materials; and maintenance cycle data reflects the recovery or impact of maintenance activities on equipment performance. This data provides comprehensive input for subsequent evaluation of the dynamic evolution rate of parameters.

[0074] Furthermore, based on the monitored equipment aging status, production batch information, and maintenance cycle data, the dynamic evolution rate of the parameters of the nonlinear influence mode is evaluated. This evolution rate reflects how quickly the parameters of the nonlinear influence mode change over time. For example, when the equipment is in a rapid aging phase or after key components have been replaced, the evolution rate of the parameters may be higher; while during stable operation, the evolution rate may be lower.

[0075] Building upon this, this application introduces an adaptive parameter update mechanism. Specifically, when the evaluated dynamic evolution rate exceeds a preset rate threshold, it indicates that the parameters of the nonlinear influence mode are changing rapidly. In this case, the system immediately triggers an update operation for the parameters of the nonlinear influence mode and adjusts the update magnitude of the parameters accordingly based on the higher dynamic evolution rate, enabling it to adapt to the current changes more quickly. Conversely, when the dynamic evolution rate is lower than the preset rate threshold, it indicates that the parameters of the nonlinear influence mode are relatively stable. In this case, the system extends the update cycle of the parameters of the nonlinear influence mode to reduce unnecessary computational resource consumption and decrease the update magnitude of the parameters to avoid over-adjustment.

[0076] Furthermore, when updating the parameters of nonlinear influence models, a weighted average method is used to fuse the parameters of historical nonlinear influence models and the parameters of the most recently calculated nonlinear influence models. This fusion mechanism aims to balance the stability of historical data and the real-time nature of the latest data. Specifically, the weighted average method dynamically adjusts the weights based on the dynamic evolution rate and data confidence level. For example, when the dynamic evolution rate is high or the confidence level of the most recently calculated data is high, the parameters of the most recently calculated nonlinear influence model will be given greater weight; conversely, the parameters of historical nonlinear influence models will receive greater weight to maintain system stability. Data confidence level can be evaluated based on factors such as the quality of data acquisition, sample size, or model fitting error.

[0077] The proposed solution continuously monitors key operational status information of the equipment and adaptively adjusts the parameter update strategy of the nonlinear influence mode accordingly, thereby ensuring the timeliness and effectiveness of parameter updates. By dynamically evaluating the parameter evolution rate, the system can intelligently determine when and by what magnitude to update, avoiding the lag or redundancy that may result from fixed-period updates. Simultaneously, a weighted average is used to fuse historical and latest parameters, and the weights are adjusted according to the dynamic evolution rate and data confidence level. This ensures that the updated parameters reflect the current true state of the system while preserving the stability of historical data, thereby improving the modeling accuracy of the relationship between local temperature, humidity, lubricant state, and instantaneous motion damping, and ultimately enhancing the prediction accuracy of motion damping on path micro-segments.

[0078] In some preferred embodiments, it is assumed that after long-term operation, the linear guides and ball screws of a pick-and-place machine experience increased friction due to wear, causing significant changes in the parameters of the nonlinear influence mode between local temperature and instantaneous motion damping. The system continuously monitors the aging state of the equipment and assesses that the dynamic evolution rate of the nonlinear influence mode parameters has exceeded a preset rate threshold. At this point, the control method of this application immediately triggers an update of the nonlinear influence mode parameters and increases the update magnitude based on the higher evolution rate. During the update process, the system uses a weighted average method to fuse the parameters of historical nonlinear influence modes with the parameters of the latest calculated nonlinear influence mode. Due to the higher dynamic evolution rate, the parameters of the latest calculated nonlinear influence mode are given a greater weight to quickly adapt to the current aging state of the equipment. For example, if the weight of the historical parameters is 0.3 and the weight of the latest parameters is 0.7, the fused parameters will be more likely to reflect the latest equipment performance. In this way, even if the equipment is in a rapid aging phase, the system can quickly adjust its prediction model for motion damping to ensure that the calculation of movement time cost remains accurate.

[0079] For example, after a comprehensive maintenance of a pick-and-place machine, the performance of its motion system recovers to near-new condition. At this point, the dynamic evolution rate of the parameters of the nonlinear influence mode will significantly decrease, even falling below a preset threshold. According to the scheme in this application, the system will extend the update cycle of the parameters of the nonlinear influence mode and reduce the update magnitude. During parameter fusion, due to the low dynamic evolution rate and stable equipment operation, the parameters of historical nonlinear influence modes may be assigned relatively large weights, while the weights of the parameters of the most recently calculated nonlinear influence mode will be relatively small, in order to maintain model stability and avoid frequent adjustments due to minor fluctuations. For example, the weight of historical parameters is 0.6, and the weight of the most recent parameters is 0.4. This strategy ensures the accuracy of predictions while avoiding unnecessary computational overhead, thus improving the system's operating efficiency.

[0080] The method further includes: During the operation of the pick-and-place machine, the generation time of parameters of historical nonlinear influence modes is continuously monitored; Based on the generation time of the parameters of the historical nonlinear influence model, calculate the time-dependent decay coefficient of the parameters of the historical nonlinear influence model. The weights of the parameters of historical nonlinear influence modes and the weights of the parameters of the latest calculated nonlinear influence modes are dynamically adjusted based on the dynamic evolution rate, data confidence, and timeliness decay coefficient. The parameters of the historical nonlinear influence model are multiplied by their corresponding weights, and the parameters of the latest calculated nonlinear influence model are multiplied by their corresponding weights. The two products are then added together to obtain the parameters of the fused nonlinear influence model.

[0081] Specifically, during the operation of the pick-and-place machine, the generation time of parameters for continuously monitoring historical nonlinear influence modes refers to the specific timestamp at which the parameters of each historical nonlinear influence mode are calculated or stored, recorded in real time by the system. This timestamp can be accurate to the second or millisecond level for subsequent timeliness assessment. The parameters of the historical nonlinear influence modes refer to the set of parameters determined at a previous point in time based on collected data and identified nonlinear influence modes.

[0082] The calculation of the timeliness decay coefficient of historical nonlinear influence model parameters, based on their generation time, refers to quantifying the "freshness" or "relevance" of historical parameters using a pre-defined decay function or model, based on the time difference between the current time and the historical parameter generation time. For example, an exponential decay function can be used, where the decay coefficient decreases as the time difference increases, or a piecewise linear decay function can be used, setting different decay rates for different time periods. The value of this timeliness decay coefficient typically ranges from 0 to 1, where 1 represents complete timeliness and 0 represents complete failure.

[0083] In practical applications, dynamically adjusting the weights of parameters from historical and newly calculated nonlinear influence models based on dynamic evolution rate, data confidence level, and time-sensitivity decay coefficient means that parameter fusion no longer relies solely on dynamic evolution rate and data confidence level, but also considers the time-sensitivity decay coefficient as an important factor. Specifically, a comprehensive weight calculation formula can be designed. For example, the weight of historical parameters can be determined by the product or weighted combination of dynamic evolution rate, data confidence level, and time-sensitivity decay coefficient, while the weight of newly calculated parameters complements the weight of historical parameters. When the time-sensitivity decay coefficient of a historical parameter is low, its weight will be correspondingly reduced, thereby minimizing its impact on the fusion result.

[0084] Furthermore, the parameters of the historical nonlinear influence model are multiplied by their corresponding weights, and the parameters of the most recently calculated nonlinear influence model are multiplied by their corresponding weights. The two products are then added together to obtain the parameters of the fused nonlinear influence model. This means that after determining the weights of the historical and latest parameters, the fusion is performed according to the principle of weighted averaging. For example, if the historical parameter is P_hist, the latest parameter is P_new, and the corresponding weights are W_hist and W_new respectively, then the fused parameter P_fused = W_hist * P_hist + W_new * P_new. The sum of W_hist and W_new is usually 1 to ensure the reasonableness of the fusion result.

[0085] In some preferred embodiments, it is assumed that the system needs to update the parameters of the nonlinear influence mode during the operation of the pick-and-place machine. The current time is T_current. There exists a parameter P_hist for a historical nonlinear influence mode, whose generation time is T_hist. The parameter of the most recently calculated nonlinear influence mode is P_new.

[0086] First, the system continuously monitors the generation time T_hist of P_hist.

[0087] Next, calculate the time-related decay coefficient of P_hist. For example, an exponential decay function can be defined: Time-related decay coefficient = exp(-(T_current-T_hist) / τ) Here, τ is a time constant representing the decay rate of parameter validity. If T_current - T_hist = 0, the decay coefficient is 1; if the time difference is large, the decay coefficient approaches 0.

[0088] Meanwhile, the system evaluates the dynamic evolution rate of the parameters of the nonlinear influence mode based on the equipment aging status, production batch information, and maintenance cycle data, and obtains the data confidence level based on the data quality assessment.

[0089] Assume the dynamic evolution rate is R and the data confidence level is C.

[0090] Then, the weights W_hist of the parameters of the historical nonlinear influence model and the weights W_new of the parameters of the most recently calculated nonlinear influence model are dynamically adjusted. For example, this can be done in the following way: W_hist=f(R, C, time-related decay coefficient) W_new=1-W_hist Here, f is a comprehensive weighting function that ensures that when R is high (rapid changes), C is low (unreliable data), or the timeliness decay coefficient is low (outdated data), W_hist decreases and W_new increases, thus relying more on the latest data. Conversely, when R is low (slow changes), C is high (reliable data), and the timeliness decay coefficient is high (timely data), W_hist increases and W_new decreases, to fully utilize the stability of historical data.

[0091] Finally, the parameters P_hist of the historical nonlinear influence model are multiplied by their corresponding weights W_hist, and the parameters P_new of the latest calculated nonlinear influence model are multiplied by their corresponding weights W_new. The two products are then added together to obtain the parameters P_fused of the fused nonlinear influence model: P_fused = W_hist * P_hist + W_new * P_new.

[0092] In this way, the fused parameter P_fused can more accurately reflect the actual operating status of the current pick-and-place machine, thereby improving the accuracy of subsequent motion damping prediction.

[0093] refer to Figure 2 An SMT placement sequence control system, applied to an SMT placement sequence control method, the system comprising: The calculation module calculates the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost for each candidate component to be placed. The synthesis module combines the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate components to be placed, and obtains the total cost of each candidate component to be placed. The sequence generation module iteratively selects the candidate component to be placed with the lowest total cost as the next placement target based on the total cost, until all candidate components to be placed are placed, generating the globally optimal placement sequence. The instruction generation module generates placement instructions that can be executed by the pick-and-place machine based on the globally optimal placement order.

[0094] Specifically, the calculation module is configured to perform cost calculations for each candidate component to be mounted. This module can acquire and process data related to component movement, potential physical interference, and solder paste adhesion degradation, and accurately calculate the corresponding movement time cost, physical interference risk cost, and solder paste adhesion degradation risk cost based on a preset algorithm model. For example, this module can integrate sensor data interfaces, data processing units, and cost calculation algorithm libraries to support real-time or near real-time cost assessment.

[0095] The synthesis module is configured to comprehensively process the various costs output by the calculation module. This module receives the costs of movement time, physical interference risks, and solder paste adhesion degradation risks, and performs weighted summation or other forms of synthesis calculations based on preset weights or a synthesis model to generate a unified total cost for each candidate component to be placed. This synthesis process aims to balance the impact of different cost factors on the placement sequence decision, ensuring the comprehensiveness of the decision.

[0096] The placement sequence generation module is configured to determine the optimal placement order based on total cost. This module receives the total cost of each candidate component from the synthesis module and employs an iterative optimization algorithm—such as a greedy algorithm, a heuristic algorithm, or a more complex one—to select the component with the lowest total cost at each step as the next placement target. This process continues until all components have been assigned placement positions, ultimately generating a globally optimal placement order. The core of this module lies in its optimization algorithm, which aims to minimize the total cost of the overall placement process.

[0097] The instruction generation module is configured to convert the generated globally optimal placement sequence into instructions executable by the pick-and-place machine. This module receives the globally optimal placement sequence output by the sequence generation module and, according to the specific pick-and-place machine's communication protocol and instruction format, converts it into a series of specific placement action instructions, such as pick-up, movement, placement, pressure reduction speed, and holding pressure time. These instructions are then sent to the pick-and-place machine control system to guide the machine in performing the actual component placement operations in the optimal sequence.

[0098] The content disclosed above is only a preferred and feasible embodiment of the present invention, and is not intended to limit the scope of protection of the present invention. Therefore, all equivalent technical changes made based on the content of the present invention specification and drawings are included within the scope of protection of the present invention. Furthermore, the elements therein can be updated as technology develops.

Claims

1. A method for controlling the SMT placement sequence, characterized in that, The method includes the following steps: For each candidate component to be placed, calculate the moving time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate component to be placed. The total cost of each candidate component to be placed is obtained by combining the moving time cost, physical interference risk cost, and solder paste adhesion decay risk cost. Based on the total cost, the candidate component to be placed with the lowest total cost is iteratively selected as the next placement target until all candidate components to be placed are placed, generating the globally optimal placement order; Based on the globally optimal placement order, placement instructions that can be executed by the pick-and-place machine are generated.

2. The SMT placement sequence control method as described in claim 1, characterized in that, The steps for calculating the cost of physical interference risk include the following: During the movement of the placement head, local temperature and humidity data of the area where the placement operation is about to be performed or the area where small components have been placed are acquired; Based on local temperature and humidity data, the predicted viscosity of solder paste on pads where small components have been placed is adjusted. When assessing the physical interference risk cost of placing large components on surrounding small components, the actual holding force of the small components is calculated based on the revised predicted solder paste viscosity, the contact area of ​​the small components, and the weight of the small components. Calculate the holding force vulnerability coefficient based on the actual holding force of the small components that have been placed. By incorporating the holding force vulnerability coefficient into the calculation of physical disturbance risk cost, the corrected physical disturbance risk cost is obtained. When the revised physical interference risk cost exceeds the preset risk threshold, adjust the pressing speed, pressing height, and holding time of large components.

3. The SMT placement sequence control method as described in claim 1, characterized in that, The steps to synthesize the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost of candidate components to obtain the total cost of each candidate component to be placed include: Set production targets; Adjust the weighting factor of the movement time cost of candidate components to be mounted according to production targets; Adjust the weighting coefficient of physical interference risk costs according to production targets; Adjust the weighting coefficient of solder paste viscosity degradation risk cost according to production targets; The adjusted weighting coefficients for the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost of the candidate components to be mounted are applied to the comprehensive calculation of the movement time cost, physical interference risk cost, and solder paste viscosity degradation risk cost of the candidate components to be mounted, so as to obtain the total cost of each candidate component to be mounted.

4. The SMT placement sequence control method as described in claim 1, characterized in that, The steps for calculating the movement time cost of candidate components to be placed include: Temperature sensors are deployed in key areas of the pick-and-place machine's motion system to collect local temperature data; Collect drive signals and position feedback data during the movement of the placement head; Based on the drive signal and position feedback data, calculate the instantaneous motion damping of the mounting head on the micro-segment of the path; Establish the relationship between local temperature and instantaneous motion damping; Based on the relationship between local temperature and instantaneous motion damping, predict the motion damping on the micro-segment of the path; Calculate the actual time for the mounting head to pass through the micro-segment of the path based on the predicted motion damping; The actual time taken for the placement head to traverse a micro-segment of the path is summed to obtain the movement time cost of the candidate components to be placed.

5. The SMT placement sequence control method as described in claim 4, characterized in that, The steps for deploying temperature sensors in key areas of the pick-and-place machine's motion system and collecting local temperature data include: The deployment location of the temperature sensor is determined based on the sensitivity of the moving parts to the effects of thermal expansion and contraction. The number of temperature sensors to be deployed is determined based on the placement head's movement frequency and dwell time. Based on the determined number of temperature sensors to be deployed, multiple temperature sensors are deployed in key areas of the linear guide rail, ball screw, and drive motor of the XY motion system of the pick-and-place machine to collect local temperature data. The deployment location of each temperature sensor is mapped to the area of ​​the motion path it monitors, so as to establish a correspondence between the motion path area and the local temperature data collected by the temperature sensor. When the mounting head moves to a specific path segment, it calls the data collected by the temperature sensor associated with that path segment according to the established correspondence, evaluates the local temperature gradient of that path segment, and uses it to predict the motion damping on the micro-segment of the path.

6. The SMT placement sequence control method as described in claim 4, characterized in that, The steps to establish the relationship between local temperature and instantaneous motion damping include: Temperature sensors, humidity sensors, and lubricant status sensors are deployed in key areas of the pick-and-place machine's motion system to collect local temperature, humidity, and lubricant status data. Collect drive signals and position feedback data during the movement of the placement head; Based on the drive signal and position feedback data, calculate the instantaneous motion damping of the mounting head on the micro-segment of the path; The local temperature, humidity, and lubricant status data are matched with instantaneous motion damping using timestamps and spatial location correlation. Based on the correlation results, the nonlinear influence patterns of local temperature, humidity, and lubricant status on instantaneous motion damping are identified to improve the accuracy of motion damping on subsequent predicted path micro-segments. Based on the nonlinear influence mode, the relationship parameters between local temperature and instantaneous motion damping are dynamically adjusted.

7. The SMT placement sequence control method as described in claim 6, characterized in that, Based on the association results, the method also includes the following steps: Based on the correlation results, the dynamic evolution trend of the nonlinear influence mode of equipment aging, production batch differences or maintenance cycle on local temperature, humidity, lubricant status and instantaneous motion damping is analyzed. Based on the dynamic evolution trend, the parameters of the nonlinear influence mode are periodically updated to adapt to the complex relationship that evolves over time, thereby improving the accuracy of motion damping on subsequent predicted path micro-segments.

8. The SMT placement sequence control method as described in claim 7, characterized in that, The steps for periodically updating the parameters of the nonlinear influence mode based on the dynamic evolution trend include: During the operation of the pick-and-place machine, continuously monitor the equipment aging status, production batch information, and maintenance cycle data; Based on equipment aging status, production batch information, and maintenance cycle data, assess the dynamic evolution rate of parameters of nonlinear influence modes. When the dynamic evolution rate exceeds the preset rate threshold, the parameters of the nonlinear influence mode are updated, and the update magnitude of the parameters of the nonlinear influence mode is adjusted according to the dynamic evolution rate. When the dynamic evolution rate is lower than the preset rate threshold, the update cycle of the parameters of the nonlinear influence mode is extended and the update amplitude of the parameters of the nonlinear influence mode is reduced. When updating the parameters of the nonlinear influence model, a weighted average method is used to fuse the parameters of the historical nonlinear influence model and the parameters of the latest calculated nonlinear influence model. The weighted average method dynamically adjusts the weights based on the dynamic evolution rate and data confidence.

9. The SMT placement sequence control method as described in claim 8, characterized in that, The method further includes: During the operation of the pick-and-place machine, the generation time of parameters of historical nonlinear influence modes is continuously monitored; Based on the generation time of the parameters of the historical nonlinear influence model, calculate the time-dependent decay coefficient of the parameters of the historical nonlinear influence model. The weights of the parameters of historical nonlinear influence modes and the weights of the parameters of the latest calculated nonlinear influence modes are dynamically adjusted based on the dynamic evolution rate, data confidence, and timeliness decay coefficient. The parameters of the historical nonlinear influence model are multiplied by their corresponding weights, and the parameters of the latest calculated nonlinear influence model are multiplied by their corresponding weights. The two products are then added together to obtain the parameters of the fused nonlinear influence model.

10. An SMT placement sequence control system, applied to the SMT placement sequence control method as described in claim 1, characterized in that, The system includes: The calculation module calculates the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost for each candidate component to be placed. The synthesis module combines the movement time cost, physical interference risk cost, and solder paste adhesion decay risk cost of the candidate components to be placed, and obtains the total cost of each candidate component to be placed. The sequence generation module iteratively selects the candidate component to be placed with the lowest total cost as the next placement target based on the total cost, until all candidate components to be placed are placed, generating the globally optimal placement sequence. The instruction generation module generates placement instructions that can be executed by the pick-and-place machine based on the globally optimal placement order.

Images