A method and system for monitoring the operating status of reflow soldering equipment

Abstract

This application relates to the field of data processing technology and discloses a method and system for monitoring the operating status of reflow soldering equipment. This method includes: collecting data through multi-point sensors, extracting temperature gradients, tension fluctuations, and gas concentration characteristics, constructing a state recognition model, calculating a health index and setting a warning threshold, constructing a fault precursor extraction model to predict future states, and finally establishing a maintenance strategy optimization system based on the fault warning information to generate an equipment maintenance plan. Through multi-dimensional data fusion and intelligent analysis, this application achieves early and accurate identification of equipment abnormalities, accurate prediction of fault development trends, and proactive maintenance decisions based on quality impact and cost optimization, thereby improving welding quality stability, reducing unplanned downtime, and lowering maintenance costs.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and system for monitoring the operating status of reflow soldering equipment. Background Art

[0002] Reflow soldering is a key process for surface mount technology (SMT) in the modern electronics manufacturing industry and is widely used in printed circuit board (PCB) assembly lines. Traditional reflow soldering equipment operating status monitoring relies primarily on a single-point alarm mechanism triggered by a preset threshold. Operators use their experience to determine equipment status and perform maintenance. As electronic products move toward higher density and miniaturization, soldering precision and quality requirements continue to increase. Some monitoring systems have begun to use simple data acquisition and statistical analysis methods, such as mean-standard deviation discrimination and trend chart analysis, to monitor key equipment parameters in real time. Some advanced systems have introduced rule-based expert systems, which use predefined if-then rules to determine abnormal conditions and combine them with regular maintenance plans to perform equipment maintenance.

[0003] However, existing reflow soldering equipment monitoring technologies have significant shortcomings. First, single-point parameter monitoring cannot fully reflect the coordinated working state of the equipment's complex thermodynamic, mechanical, and gas systems, resulting in inaccurate or delayed anomaly detection. Second, methods with preset fixed thresholds are difficult to adapt to different production processes and environmental conditions, resulting in false positives or missed reports. Third, simple statistical analysis methods cannot identify the complex correlations between parameters and the evolution of faults, lacking the ability to predict fault development trends. Fourth, existing maintenance strategies are mostly passive responses or fixed cycles, failing to comprehensively optimize equipment status, production plans, and maintenance costs, resulting in unnecessary downtime losses or premature or late maintenance interventions. Finally, maintenance decisions lack a direct connection to product quality, making it difficult to strike a balance between quality assurance and cost control. Summary of the Invention

[0004] The present application provides a method and system for monitoring the operating status of reflow soldering equipment, which is used to achieve early and accurate identification of equipment abnormal conditions, accurate prediction of fault development trends, and proactive maintenance decisions based on quality impact and cost optimization through multi-dimensional data fusion and intelligent analysis, thereby improving soldering quality stability, reducing unplanned downtime and lowering maintenance costs.

[0005] In a first aspect, the present application provides a method for monitoring the operating status of reflow soldering equipment, the method comprising: collecting operating parameters through multi-point sensors of the reflow soldering equipment to obtain a welding equipment status data set; extracting a temperature gradient distribution map, conveyor belt tension fluctuation characteristics and a gas concentration time-varying curve based on the welding equipment status data set, and performing a multi-dimensional feature cross-correlation analysis to obtain a welding equipment operating feature matrix; constructing a state recognition model based on the welding equipment operating feature matrix, classifying the equipment operating status through a dynamic threshold adaptive adjustment mechanism to obtain a welding equipment status category; calculating an equipment health index based on the welding equipment status category, and setting a multi-level warning threshold to obtain a welding equipment health assessment result; utilizing the welding equipment health assessment result to construct a fault precursor feature extraction and amplification model, predicting the future state by comparing the reflow soldering feature time series pattern library, and obtaining welding equipment fault warning information; constructing a multi-objective maintenance strategy optimization system based on the welding equipment fault warning information, generating a maintenance plan through correlation analysis between welding quality and equipment status, and obtaining a welding equipment maintenance plan.

[0006] In a second aspect, the present application provides a reflow soldering equipment operating status monitoring system, the reflow soldering equipment operating status monitoring system comprising:

[0007] An acquisition module is used to collect operating parameters of the reflow soldering equipment through multi-point sensors to obtain a soldering equipment status data set;

[0008] a correlation module for extracting a temperature gradient distribution map, conveyor belt tension fluctuation characteristics, and a gas concentration time-varying curve based on the welding equipment status data set, and performing a multi-dimensional feature cross-correlation analysis to obtain a welding equipment operation feature matrix;

[0009] A construction module is used to construct a state recognition model based on the welding equipment operation feature matrix, classify the equipment operation state through a dynamic threshold adaptive adjustment mechanism, and obtain a welding equipment state category;

[0010] A calculation module, configured to calculate an equipment health index according to the welding equipment status category, and set a multi-level warning threshold to obtain a welding equipment health assessment result;

[0011] A comparison module is used to construct a fault precursor feature extraction and amplification model using the welding equipment health assessment results, predict future states by comparing with the reflow soldering feature timing pattern library, and obtain welding equipment fault warning information;

[0012] A generation module is used to build a multi-objective maintenance strategy optimization system based on the welding equipment fault warning information, generate a maintenance plan through correlation analysis between welding quality and equipment status, and obtain a welding equipment maintenance plan.

[0013] The third aspect of the present invention provides a computer device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the computer device executes the above-mentioned reflow soldering equipment operation status monitoring method.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned method for monitoring the operating status of reflow soldering equipment.

[0015] In the technical solution provided by this application, a comprehensive welding equipment status data set is obtained by collecting operating parameters through multi-point sensors, realizing comprehensive monitoring of multiple systems and multiple parameters of the equipment. Compared with the traditional single-point monitoring method, the comprehensiveness and accuracy of data collection are greatly improved, providing a rich data foundation for subsequent analysis; the temperature gradient distribution map, conveyor belt tension fluctuation characteristics and gas concentration time-varying curve are extracted, and multi-dimensional feature cross-correlation analysis is performed to obtain the welding equipment operation feature matrix. This process fully explores the inherent correlation between multi-dimensional parameters and reveals the mutual influence mechanism between different subsystems. The abnormal state detection is no longer limited to the judgment of a single parameter exceeding the limit, but can be evaluated from the overall performance of the system; the state recognition model constructed based on the welding equipment operation feature matrix realizes accurate classification of the equipment operation state through a dynamic threshold adaptive adjustment mechanism. This mechanism dynamically adjusts the classification threshold according to the state transition frequency and duration in the equipment operation history data, effectively solving the misjudgment problem caused by fixed thresholds and improving the accuracy and stability of state recognition. In particular, in the combined application of the three machine learning algorithms of support vector machine, random forest and gradient boosting decision tree, the respective advantages of different algorithms are fully utilized. The support vector machine is good at processing high-dimensional features. The random forest algorithm has good noise resistance and feature importance assessment capabilities, while the gradient boosting decision tree is good at handling unbalanced data and capturing subtle feature changes. The weighted fusion of the three algorithms significantly improves classification performance. The equipment health index is calculated based on the welding equipment status category, and multi-level warning thresholds are set, achieving a transition from status classification to health quantification, making equipment status assessment more precise and intuitive. The welding equipment health assessment results are used to construct a fault precursor feature extraction and amplification model. The future status is predicted by comparing the reflow soldering feature time series pattern library, transforming passive response into active prediction, significantly improving fault warning time and gaining sufficient preparation time for maintenance decision-making. A multi-objective maintenance strategy optimization system is constructed based on welding equipment fault warning information. Maintenance plans are generated by analyzing the correlation between welding quality and equipment status. For the first time, product quality factors are introduced into the maintenance decision-making process, achieving a transition from purely equipment-oriented to coordinated optimization of quality and cost. The system solves the optimal maintenance time window through an integer programming algorithm, minimizing total maintenance cost while ensuring welding quality. This provides scientific and economical maintenance decision support for enterprises, significantly improving production efficiency and product yield, and reducing maintenance costs and unplanned downtime losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A schematic diagram of an embodiment of a method for monitoring the operating status of a reflow soldering device in an embodiment of the present application;

[0018] Figure 2 This is a schematic diagram of an embodiment of a reflow soldering equipment operating status monitoring system in an embodiment of the present application;

[0019] Figure 3 It is a schematic block diagram of the structure of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The embodiments of the present application provide a method and system for monitoring the operating status of a reflow soldering device. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or devices.

[0021] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In one embodiment of the present application, a method for monitoring the operating status of a reflow soldering device includes:

[0022] Step S101: Collect operating parameters of the reflow soldering equipment through a multi-point sensor to obtain a soldering equipment status data set;

[0023] Step S102: extracting the temperature gradient distribution map, conveyor belt tension fluctuation characteristics, and gas concentration time-varying curve based on the welding equipment status data set, and performing multi-dimensional feature cross-correlation analysis to obtain a welding equipment operation feature matrix;

[0024] Step S103: constructing a state recognition model based on the welding equipment operation feature matrix, classifying the equipment operation state through a dynamic threshold adaptive adjustment mechanism, and obtaining a welding equipment state category;

[0025] Step S104: Calculate the equipment health index according to the welding equipment status category, set multi-level warning thresholds, and obtain the welding equipment health assessment result;

[0026] Step S105: construct a fault precursor feature extraction and amplification model using the welding equipment health assessment results, predict the future state by comparing the reflow soldering feature timing pattern library, and obtain welding equipment fault warning information;

[0027] Step S106: construct a multi-objective maintenance strategy optimization system based on the welding equipment fault warning information, generate a maintenance plan through correlation analysis between welding quality and equipment status, and obtain a welding equipment maintenance plan.

[0028] It is understandable that the execution subject of the present application can be a reflow soldering equipment operating status monitoring system, or a terminal or a server, which is not limited here. The embodiment of the present application is described by taking the server as the execution subject as an example.

[0029] Specifically, multi-point sensor acquisition of operating parameters begins with the installation of temperature sensor arrays in the preheating, reflow, and cooling zones of the reflow soldering equipment. These arrays consist of a grid-like structure of multiple PT100 platinum resistance temperature sensors arranged longitudinally and transversely along the soldering furnace. Speed ​​sensors and tension sensors are installed in the conveyor system. The speed sensor uses a photoelectric encoder, while the tension sensor uses a strain gauge force sensor. The gas supply system is equipped with gas concentration sensors and pressure sensors. The gas concentration sensor monitors nitrogen or shielding gas concentration, while the pressure sensor monitors gas delivery pressure. The power supply system is equipped with current and voltage sensors to monitor power parameters in each heating zone. All sensors are sampled uniformly through an industrial IoT module. The sampling frequency for temperature data is 10 Hz, for conveyor parameters is 5 Hz, for gas parameters is 2 Hz, and for electrical parameters is 20 Hz. The collected raw data requires preprocessing. A sliding average filter algorithm is used to denoise the temperature data. This algorithm smooths out short-term fluctuations by taking the average of multiple consecutive sampling points, replacing the current value. A median filter algorithm is applied to conveyor belt speed data. This algorithm eliminates sporadic interference by taking the median value of the data within a time window. A Kalman filter algorithm is applied to gas concentration data. This algorithm uses an iterative prediction-correction process to eliminate noise and improve signal stability. When extracting features and performing multidimensional feature cross-correlation analysis, the preprocessed data is first segmented into time windows. The window length is adaptively adjusted to 5 to 30 minutes based on the equipment operating cycle. A temperature gradient distribution map is calculated for the temperature data within each time window. The spatial temperature gradient is calculated by dividing the difference between adjacent temperature sensors by the sensor spacing, forming a temperature gradient spatiotemporal distribution matrix. The temperature gradient distribution map reflects the uniformity and stability of heat transfer during welding and is crucial to weld quality. Tension fluctuation characteristics are extracted from conveyor belt parameters, including peak-to-peak tension, tension stability coefficient, and tension pulsation frequency. Peak-to-peak tension refers to the difference between the maximum and minimum tension values. The tension stability coefficient is the ratio of the tension standard deviation to the mean. The tension pulsation frequency is the dominant frequency component obtained by performing a fast Fourier transform on the tension time series data. The gas concentration data was transformed into a time-frequency form to extract the characteristics of the gas concentration time-varying curve, including the gas concentration change rate, gas concentration fluctuation period, and the proportion of the gas concentration stable interval. These characteristics were then subjected to dimensionality reduction using principal component analysis, retaining the principal components with a cumulative contribution rate of 95% to form a welding equipment operation characteristic matrix.

[0030] The state recognition model is constructed by first predefining five typical operating states of reflow soldering equipment: normal operation, abnormal temperature, abnormal conveyor belt, abnormal gas system, and abnormal electrical system. The welding equipment operational feature matrix is ​​divided into a training matrix and a validation matrix, typically with a ratio of 70% to 30%. Three different classifiers are constructed using the support vector machine algorithm, the random forest algorithm, and the gradient boosting decision tree algorithm, respectively. The support vector machine algorithm classifies states by finding the optimal classification hyperplane, the random forest algorithm determines the state type by voting among multiple decision trees, and the gradient boosting decision tree algorithm improves classification performance by iteratively training a series of weak classifiers and then performing a weighted combination. A dynamic threshold adaptive adjustment mechanism is constructed to dynamically adjust the discrimination threshold parameters of the three classifiers based on the state transition frequency and duration in historical equipment operation data, thereby improving classification accuracy. Finally, the discrimination results of the three classifiers are fused using a weighted voting method to obtain the final welding equipment state category. The health index calculation method first converts the welding equipment status category into a quantified subsystem status value: normal operation is assigned a value of 1.0, mild abnormality is assigned a value of 0.8, moderate abnormality is assigned a value of 0.6, and severe abnormality is assigned a value of 0.3. Subsystem weight coefficients are assigned based on the degree of impact of each subsystem on the overall equipment performance. Typically, the temperature system is assigned a weight of 0.4, the conveyor system a weight of 0.3, the gas system a weight of 0.2, and the electrical system a weight of 0.1. The quantified subsystem status value is multiplied by the subsystem weight coefficient to obtain a weighted health score. The abnormality duration characteristic is extracted from the historical changes in the welding equipment status category to obtain the abnormality persistence factor. Based on the correlation between multiple subsystem abnormalities within the welding equipment status category, the abnormality diffusion factor is quantified to determine the degree of abnormality spread within the system. The final equipment health index is calculated by combining the weighted status score, abnormality persistence factor, and abnormality diffusion factor. Multi-level warning thresholds are then set. Typically, a health index below 0.9 is considered a mild warning, below 0.7 is considered a moderate warning, and below 0.5 is considered a severe warning.

[0031] To extract fault precursor features and predict future states, the welding equipment health assessment results are first correlated with equipment parameters at historical time points to construct a time series mapping database. Fault evolution patterns are extracted from the time series mapping database, and the characteristic changes preceding a decline in the equipment health index are identified to generate a set of fault precursor features. Weak changes in the fault precursor feature set are nonlinearly amplified, typically with a factor of 2 to 5, to make subtle trend changes easier to identify. Typical evolution paths and feature sequences for different types of faults are collected to form a reflow soldering feature time series pattern library. Similarity calculations are performed between the amplified fault precursor features and the reflow soldering feature time series pattern library. Euclidean distance or DTW (Dynamic Time Warping) algorithms are used to calculate feature sequence similarity and identify potential fault types and development trends. Based on the fault matching score and the equipment health index change trend, future equipment states and potential failure risks are predicted. To construct a maintenance strategy optimization system, fault warning information for welding equipment is mapped and correlated with the equipment's physical structure and operating logic. Correlations between historical welding quality data and equipment status parameters are analyzed to generate a quality impact factor matrix. For different maintenance strategies, equipment downtime losses, maintenance labor costs, and spare parts costs are calculated to generate a maintenance cost estimate table. Using the quality impact factor matrix and the maintenance cost estimate table as constraints, an integer programming algorithm is used to calculate the optimal maintenance time window. Based on the fault type and equipment structure information from the fault warning information, specific maintenance items and operation procedures are generated, forming a complete maintenance plan that includes maintenance time, maintenance items, required personnel, and resource allocation.

[0032] For example, during an actual monitoring process, after 24 hours of continuous operation, the temperature data from the mid-reflow zone of a reflow soldering machine's temperature sensor array fluctuated by 3°C. The temperature gradient between adjacent sensors increased from the standard 5°C / cm to 7.2°C / cm. Simultaneously, the tension stability coefficient in the conveyor belt tension fluctuation signature increased from 0.05 to 0.12, and the gas concentration change rate increased from 0.5% / min to 1.2% / min. Through feature extraction and cross-correlation analysis, the resulting operating feature matrix was identified by the state recognition model as a temperature anomaly. The equipment health index dropped from 0.95 to 0.82, triggering a minor warning. The fault precursor feature extraction model zoomed in on the temperature fluctuation signature and compared it with a time series pattern library, revealing an 85% similarity with a heating element aging failure mode. The system predicted that heating element power instability would occur within the next 48 hours. Based on this warning information, the maintenance strategy optimization system generated a maintenance plan, recommending replacement of the relevant heating element after the end of that day's production and adjusting the reflow zone temperature control parameters to avoid unplanned downtime during peak production periods, minimize production disruptions, and ensure soldering quality.

[0033] In the embodiment of the present application, a comprehensive welding equipment status data set is obtained by collecting operating parameters through multi-point sensors, realizing comprehensive monitoring of multiple systems and multiple parameters of the equipment. Compared with the traditional single-point monitoring method, the comprehensiveness and accuracy of data collection are greatly improved, providing a rich data basis for subsequent analysis; the temperature gradient distribution map, conveyor belt tension fluctuation characteristics and gas concentration time-varying curve are extracted, and multi-dimensional feature cross-correlation analysis is performed to obtain the welding equipment operation feature matrix. This process fully explores the inherent correlation between multi-dimensional parameters and reveals the mutual influence mechanism between different subsystems, so that abnormal state detection is no longer limited to the judgment of a single parameter exceeding the limit, but can be evaluated from the overall performance of the system; the state recognition model constructed based on the welding equipment operation feature matrix realizes accurate classification of the equipment operation state through a dynamic threshold adaptive adjustment mechanism. This mechanism dynamically adjusts the classification threshold according to the state transition frequency and duration in the equipment operation history data, effectively solving the misjudgment problem caused by fixed thresholds, and improving the accuracy and stability of state recognition. In particular, in the combined application of the three machine learning algorithms of support vector machine, random forest and gradient boosting decision tree, the respective advantages of different algorithms are fully utilized. The support vector machine is good at processing high-dimensional feature space. The random forest algorithm has good noise resistance and feature importance assessment capabilities, while the gradient boosting decision tree is good at handling unbalanced data and capturing subtle feature changes. The weighted fusion of the three algorithms significantly improves classification performance. The equipment health index is calculated according to the welding equipment status category, and multi-level warning thresholds are set, achieving a transition from status classification to health quantification, making equipment status assessment more precise and intuitive. The welding equipment health assessment results are used to construct a fault precursor feature extraction and amplification model. The future state is predicted by comparing the reflow soldering feature time series pattern library, transforming passive response into active prediction, greatly improving the fault warning time and gaining sufficient preparation time for maintenance decision-making. A multi-objective maintenance strategy optimization system is constructed based on welding equipment fault warning information. Maintenance plans are generated by analyzing the correlation between welding quality and equipment status. For the first time, product quality factors are introduced into the maintenance decision-making process, achieving a transition from purely equipment-oriented to coordinated optimization of quality and cost. The system solves the optimal maintenance time window through an integer programming algorithm, minimizing the total maintenance cost while ensuring welding quality. This provides scientific and economical maintenance decision-making support for enterprises, significantly improving production efficiency and product yield, and reducing maintenance costs and unplanned downtime losses.

[0034] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0035] (1) Install temperature sensor arrays in the preheating zone, reflow zone, and cooling zone of the reflow soldering equipment, install speed sensors and tension sensors in the conveyor belt system, install gas concentration sensors and air pressure sensors in the gas supply system, and install current and voltage sensors in the power supply system to form a multi-type sensor network;

[0036] (2) The industrial Internet of Things module is used to uniformly sample the data collected by the multi-type sensor network, where the sampling frequency of temperature data is 10 Hz, the sampling frequency of conveyor parameters is 5 Hz, the sampling frequency of gas parameters is 2 Hz, and the sampling frequency of electrical parameters is 20 Hz;

[0037] (3) The collected temperature data is denoised using a sliding average filter algorithm, the conveyor belt speed data is denoised using a median filter algorithm, and the gas concentration data is processed using a Kalman filter algorithm;

[0038] (4) Missing data points are supplemented by linear interpolation or polynomial interpolation according to their continuous missing length, and outliers are preliminarily screened by the statistical 3σ principle;

[0039] (5) Align the processed sensor data according to a unified timestamp to generate a welding equipment status data set.

[0040] Specifically, the monitoring method is based on temperature sensor arrays installed in the preheating, reflow, and cooling zones of the reflow soldering equipment. These arrays consist of multiple PT100 platinum resistance temperature sensors arranged in a grid pattern, typically with 16-25 sensors installed in each zone, forming a complete temperature field monitoring network. The preheating zone typically has a temperature range of 150-180°C, the reflow zone has a temperature range of 230-260°C, and the cooling zone gradually decreases from high to room temperature. Speed ​​sensors and tension sensors are installed in the conveyor belt system. The speed sensor uses a high-precision photoelectric encoder capable of monitoring conveyor belt speed with an accuracy of 0.1 mm / s. The tension sensor uses a strain gauge force sensor with a measurement range of 0-50 N to monitor changes in conveyor belt tension. A gas concentration sensor and a pressure sensor are installed in the gas supply system. The gas concentration sensor monitors the concentration of the protective atmosphere (usually nitrogen) with a measurement range of 0-100%, while the pressure sensor has a measurement range of 0-1 MPa to monitor the pressure stability of the gas supply system. Current and voltage sensors with a measurement range of 0-100A and 0-380V are installed in the power supply system to monitor the power parameters of the heating elements. This multi-type sensor network collects unified data through an Industrial Internet of Things (IIoT) module, using industrial fieldbus protocols such as ModBus-RTU or Profibus-DP for communication, ensuring stable and reliable data transmission. Due to the varying rates of change and importance of different parameters, differentiated sampling frequencies are set: temperature data is sampled at 10Hz, collecting temperature data every 0.1 seconds to capture dynamic temperature changes; conveyor belt parameters are sampled at 5Hz, collecting speed and tension data every 0.2 seconds; gas parameters are sampled at 2Hz, collecting gas concentration and pressure data every 0.5 seconds; and electrical parameters are sampled at 20Hz, collecting current and voltage data every 0.05 seconds to monitor rapid changes in the electrical system. This differentiated sampling strategy ensures accurate monitoring of key parameters while avoiding data redundancy.

[0041] The collected raw data often contains noise and interference, requiring signal processing. Temperature data is denoised using a sliding average filter algorithm. This algorithm calculates the average of N consecutive points to replace the current value. The filter window size is typically set to 5-9 points, effectively smoothing temperature fluctuations without excessively delaying the response to temperature changes. Conveyor speed data is denoised using a median filter algorithm. This algorithm eliminates short-term outliers by taking the median value of the data within a specified time window. The window size is typically 7 points, effectively eliminating spikes in conveyor speed data. Gas concentration data is processed using a Kalman filter algorithm. This algorithm establishes a state prediction model and a measurement model for gas concentration changes, and combines the process noise covariance and measurement noise covariance for recursive optimal estimation. This algorithm effectively addresses random fluctuations and systematic errors in gas concentration data. During data acquisition, data loss due to temporary sensor failure or communication interruptions is inevitable, and these missing data must be supplemented. For data with consecutive missing values ​​less than three sampling points, linear interpolation is used to fill in the missing values, estimating the missing values ​​by linearly connecting adjacent valid data points. For data with consecutive missing values ​​between three and ten sampling points, polynomial interpolation methods such as cubic spline interpolation are used to fill in the missing values, which can maintain data smoothness and continuity. Outliers are initially screened using the statistical 3σ principle: data outside the range of ±3 standard deviations from the mean are marked as suspected outliers, and whether to replace them with interpolated values ​​is determined based on the changing trends of the preceding and following data.

[0042] Because different sensors have varying sampling frequencies, subsequent multi-sensor fusion analysis requires aligning the processed sensor data to a unified timestamp. Using the minimum common sampling period (0.1 seconds) as a benchmark, low-frequency sampling data is interpolated, maintaining the value of the previous sampling point between two sampling points to form a time-consistent data structure. This processed dataset includes the temperature field distribution, conveyor operating parameters, gas system parameters, and electrical system parameters, forming a comprehensive dataset reflecting the operating status of the welding equipment and providing a data foundation for subsequent feature extraction and state recognition.

[0043] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0044] (1) The welding equipment status data set is segmented according to different time window lengths. The window length is adaptively adjusted to 5 minutes to 30 minutes according to the operation cycle of different equipment;

[0045] (2) Calculate the temperature gradient distribution map for the temperature data in each time window. The spatial temperature gradient is obtained by taking the difference between adjacent temperature sensors and dividing it by the sensor spacing. Then, the spatial temperature gradient distribution matrix is ​​formed by combining the time dimension.

[0046] (3) Extract the belt tension fluctuation characteristics from the belt parameters, including the tension peak-to-peak value, tension stability coefficient, and tension pulsation frequency characteristics, and construct the tension fluctuation feature vector;

[0047] (4) Perform time-frequency transformation on the gas concentration data to extract the characteristics of the gas concentration time-varying curve, including the gas concentration change rate, gas concentration fluctuation period and gas concentration stable interval ratio;

[0048] (5) Calculate the phase correlation between the temperature gradient distribution map and the conveyor belt tension fluctuation characteristics, analyze the time-lag correlation between the gas concentration time-varying curve and the temperature gradient, and construct the parameter cross-influence matrix;

[0049] (6) The extracted features are reduced in dimension by principal component analysis, and the principal components with a cumulative contribution rate of 95% are retained to form the welding equipment operation feature matrix.

[0050] Specifically, processing the soldering equipment status dataset begins with time window segmentation, segmenting the continuously collected data into different time window lengths. The window length is adaptively adjusted to range from 5 to 30 minutes based on the operating cycle of the equipment. This adaptive window length is determined based on the operating characteristics of the reflow soldering equipment. For small reflow soldering equipment on a fast PCB assembly line, a production cycle may only take about 5 minutes, so a 5-minute time window is selected. For large reflow soldering equipment processing large or multi-layer PCBs, a complete soldering cycle may take up to 30 minutes. The time window selection uses autocorrelation analysis to determine the optimal window size. This involves calculating the autocorrelation coefficient of data at different time intervals. The time interval where the autocorrelation coefficient falls below 0.3 is used as the reference value for the effective window size. In practice, sliding window processing is used, with 50% overlap between adjacent windows to ensure that state changes at the window boundaries are not missed. Calculating the temperature gradient distribution map for the temperature data within each time window is a key step in understanding the soldering heat distribution. The temperature gradient distribution map is calculated by dividing the difference between adjacent temperature sensors by the sensor spacing to obtain the spatial temperature gradient. For example, if the distance between two adjacent sensors is 5 cm, and one sensor's temperature is 200°C and the other's is 215°C, the spatial temperature gradient at that point is (215 - 200°C) / 5 = 3°C / cm. The temperature field of the entire welding equipment is represented using a three-dimensional coordinate system (x, y, z), where x and y represent the horizontal and vertical position coordinates, respectively, and z represents the temperature value. The temperature gradient in each direction is calculated by calculating (∂T / ∂x) and (∂T / ∂y). Incorporating the time dimension, a spatiotemporal distribution matrix of the temperature gradient is formed. This involves obtaining a two-dimensional temperature gradient matrix at each sampling time point, which then evolves over time to form a three-dimensional matrix. The spatiotemporal distribution matrix of the temperature gradient is mathematically represented as V(x, y, t), where (x, y) represents the spatial position, t represents the time, and the matrix element value represents the temperature gradient at that position and time.

[0051] Tension fluctuation characteristics include peak-to-peak tension, tension stability coefficient, and tension pulsation frequency characteristics. The peak-to-peak tension value is calculated by finding the maximum and minimum tension values ​​within a time window. The difference between the two is the peak-to-peak value, reflecting the extreme range of tension fluctuations. The tension stability coefficient is calculated by dividing the standard deviation of the tension values ​​by the average value, indicating the degree of stability of the tension fluctuation relative to the average tension. The tension pulsation frequency characteristic is determined by performing a fast Fourier transform (FFT) on the tension time series data to determine the main frequency components and their amplitudes. For example, a 512-point FFT analysis is performed on tension data with a sampling rate of 5Hz to obtain the frequency spectrum distribution within the range of 0-2.5Hz. The top three frequencies with the largest amplitudes and their corresponding amplitudes are found to form a frequency-amplitude feature vector of length 6. These three features constitute the tension fluctuation feature vector, which is used to characterize the dynamic characteristics of the conveyor belt system. Performing time-frequency transformation on gas concentration data is a necessary step in understanding the performance of the protective gas system. First, a continuous wavelet transform (CWT) was used to perform time-frequency analysis on the gas concentration time series. Morlet wavelet was selected as the mother wavelet function. After transformation, the time-frequency spectrum of the gas concentration was obtained, showing the energy distribution at different time points and frequencies. Three types of features were extracted from the time-frequency spectrum: the gas concentration change rate was calculated by first-order differences of the original concentration data, representing the magnitude of concentration change per unit time; the gas concentration fluctuation period was determined by analyzing the frequencies corresponding to the energy peaks in the time-frequency spectrum, with the reciprocals of the three frequency components with the highest energy being taken as the main period; and the proportion of gas concentration stable intervals was determined by calculating the proportion of time points where the concentration change rate was less than a preset threshold (typically 0.5% / min). These features comprehensively reflect the stability and dynamic response characteristics of the gas system.

[0052] Calculating the phase correlation between the temperature gradient distribution and the belt tension fluctuation characteristics is an important means to understand the interaction between the temperature field and the mechanical system. The phase correlation is calculated using the following formula:

[0053]

[0054] in, The time lag is The phase correlation when is the number of sampling points in the time window, is the number of sampling points in the temperature gradient space, is the dimension of the tension eigenvector, Indicates location ( ) in time The temperature gradient value, Indicates time By calculating the tension eigenvector at different time delays The phase correlation under the condition of θ is calculated and the time lag value corresponding to the maximum correlation is found, which reflects the time relationship between temperature change and conveyor belt tension change.

[0055] Similarly, the cross-correlation function is used to analyze the time-delayed correlation between the gas concentration time-varying curve and the temperature gradient:

[0056]

[0057] in, represents the time-lagged correlation coefficient between gas concentration and average temperature gradient, It's time The gas concentration, It's time The average temperature gradient, and are the mean of the gas concentration and the mean temperature gradient, is the number of sampling points, is a time-lagged variable. The correlation coefficients under the same value are calculated, and a parameter cross-influence matrix is ​​constructed. The elements in this matrix represent the maximum correlation coefficients between parameters and their corresponding time lag values. Dimensionality reduction of the extracted features using principal component analysis is an effective means of converting high-dimensional features into low-dimensional principal components. Principal component analysis first standardizes all features so that the mean of each feature is 0 and the variance is 1. Then, the covariance matrix is ​​calculated and eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​and eigenvectors. The eigenvectors are arranged in descending order according to the corresponding eigenvalues. The contribution rate and cumulative contribution rate of each eigenvalue are calculated. The top k eigenvectors with a cumulative contribution rate of 95% are selected as the projection matrix. The original high-dimensional features are converted into k-dimensional principal components through projection transformation to form the welding equipment operation feature matrix.

[0058] Taking an actual reflow soldering process as an example, a reflow soldering machine at an electronics assembly plant soldered a batch of double-layer PCBs during production. Based on the production cycle, a 15-minute time window was set for data analysis. Within this time window, a temperature gradient distribution map was calculated using data from 10 temperature sensors in the preheating zone, 12 in the reflow zone, and 8 in the cooling zone. An abnormal temperature gradient was observed in the central reflow zone, with the maximum gradient reaching 8°C / cm, exceeding the 5°C / cm standard under normal operating conditions. Tension characteristics were also extracted from data collected from the conveyor system, yielding a peak-to-peak tension value of 4.2N and a tension stability coefficient of 0.08. FFT analysis revealed the main pulsation frequencies of 0.12Hz, 0.25Hz, and 0.43Hz, corresponding to amplitudes of 0.7N, 0.5N, and 0.3N, respectively. Time-frequency transformation of the gas concentration data revealed a concentration change rate of 0.8% / min, with major fluctuation periods of 180, 90, and 45 seconds, and a stable interval ratio of 78%. Phase correlation analysis of temperature gradients and tension fluctuations yielded a maximum correlation coefficient of 0.72, corresponding to a time lag of 8 seconds, indicating that temperature changes precede tension fluctuations by 8 seconds. Time-lag correlation analysis of gas concentration and temperature gradients revealed a maximum correlation coefficient of 0.65, corresponding to a time lag of 15 seconds, indicating that temperature gradient changes precede gas concentration changes by 15 seconds. These analysis results were integrated into a parameter cross-influence matrix and then processed through principal component analysis, reducing the original 43-dimensional feature space to an 8-dimensional principal component space. The cumulative contribution of the eight principal components reached 96.4%, effectively reducing data dimensionality while retaining key information. The resulting 8×15-dimensional welding equipment operational feature matrix served as input for subsequent state identification.

[0059] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0060] (1) Predefine five typical operating states of reflow soldering equipment: normal operating state, abnormal temperature state, abnormal conveyor belt state, abnormal gas system state and abnormal electrical system state;

[0061] (2) Divide the welding equipment operation feature matrix into a training matrix and a verification matrix. The training matrix is ​​used for subsequent classification algorithm training.

[0062] (3) Process the training matrix through the support vector machine algorithm to construct the first classifier, and use the verification matrix to evaluate the performance of the first classifier to obtain the first classification accuracy;

[0063] (4) The training matrix is ​​processed by the random forest algorithm to construct the second classifier, and the performance of the second classifier is evaluated using the validation matrix to obtain the second classification accuracy;

[0064] (5) The training matrix is ​​processed by the gradient boosting decision tree algorithm to construct a third classifier, and the performance of the third classifier is evaluated using the validation matrix to obtain the third classification accuracy;

[0065] (6) Construct a dynamic threshold adaptive adjustment mechanism to dynamically adjust the discrimination threshold parameters of the first classifier, the second classifier, and the third classifier according to the state transition frequency and duration in the equipment operation history data;

[0066] (7) According to the first classification accuracy, the second classification accuracy and the third classification accuracy, the discrimination results of the three classifiers are fused by the weighted voting method to obtain the welding equipment status category.

[0067] Specifically, five typical operating states of reflow soldering equipment are predefined as classification targets. Normal operating state refers to a state where all equipment parameters fluctuate within the set range, the temperature distribution is uniform, the conveyor belt operates smoothly, the gas concentration is stable, and the electrical parameters are normal. Abnormal temperature state refers to an abnormal temperature gradient distribution pattern, such as excessively high or low temperatures in a certain area, temperature fluctuations exceeding a set threshold, or temperature gradients exceeding a set safety range. Abnormal conveyor belt state refers to unstable conveyor speed, tension fluctuations exceeding a safety threshold, or conveyor belt jitter or slippage. Abnormal gas system state refers to large fluctuations in shielding gas concentration, unstable gas pressure, or insufficient gas supply. Abnormal electrical system state refers to abnormal current and voltage fluctuations in the heating element, power fluctuations outside the set range, or abnormal electrical control system response. Partitioning the welding equipment operating feature matrix into a training matrix and a validation matrix is ​​a fundamental step in building a machine learning model. A typical partitioning ratio is 70% of the data for training and 30% for validation. Alternatively, a K-fold cross-validation method can be used, where the data is divided equally into K parts, with K-1 parts selected each time as training data and the remaining 1 part selected as validation data. The results are averaged after K training cycles. Stratified sampling is used when partitioning the data to ensure that the proportions of various states in the training and validation sets are consistent, thus avoiding model bias caused by uneven data distribution. For reflow soldering equipment, several months of operating data are typically collected, covering a variety of typical operating conditions and abnormal conditions, to form a complete training and validation dataset.

[0068] The support vector machine algorithm is a binary classification model that classifies data by finding the optimal classification hyperplane. For the multi-class classification problem of reflow soldering equipment, a one-vs-many strategy is employed: a binary classifier is constructed for each state, with each classifier determining whether the current state belongs to a specific category. The core of the support vector machine is to find the maximum-margin hyperplane. For cases where linear separation is not possible, a kernel function is introduced to map the data into a high-dimensional space. A radial basis function (RBF) kernel is typically used for reflow soldering equipment state recognition. The kernel parameters are optimized through grid search and cross-validation. The training matrix is ​​processed using the support vector machine algorithm to construct a first classifier. Its performance is then evaluated using a validation matrix. Metrics such as accuracy, precision, recall, and F1 score are calculated to comprehensively evaluate the performance of the first classifier and determine the first classification accuracy. The random forest algorithm is an ensemble learning method that constructs multiple decision trees and uses the majority vote as the final classification result. Each decision tree in the random forest is trained on a random subset of the original training data, and each node split considers a random subset of features to increase diversity among the trees and reduce the risk of overfitting. For reflow soldering equipment status recognition, the number of decision trees is typically set to 100-500, with a maximum tree depth of 10-20 layers to balance model complexity and generalization. An advantage of random forests is that they provide feature importance scores, helping to identify the most critical feature parameters for status classification. The training matrix is ​​processed using the random forest algorithm to construct a second classifier. The validation matrix is ​​used to calculate various performance metrics and determine the second classification accuracy.

[0069] The gradient boosting decision tree algorithm trains a series of decision trees sequentially, with each new tree fitting the residuals of the previous tree, gradually improving model performance. For reflow soldering equipment state recognition, commonly implemented methods include XGBoost or LightGBM. These algorithms optimize the second-order derivative of the objective function and use regularization techniques to improve computational efficiency and model performance. Typical parameter settings include a learning rate of 0.01-0.1, a maximum tree depth of 3-7, and a number of trees of 100-1000. The training matrix is ​​processed using the gradient boosting decision tree algorithm to construct a third-class classifier. Its performance is evaluated using a validation matrix, and various performance metrics are calculated to obtain the third-class accuracy. Establishing a dynamic threshold adaptive adjustment mechanism is a key step in improving classifier robustness. The discrimination threshold parameters of each classifier are dynamically adjusted based on the frequency and duration of state transitions in historical equipment operation data. For reflow soldering equipment, state transitions are typically low, normal states last for a long time, and abnormal states are brief but require a rapid response. Dynamic threshold adjustment is based on time series analysis, calculating the average duration and transition probability of different states. This model constructs a Markov state transition model and dynamically adjusts the threshold based on the current state and duration. For example, when a device remains in a normal state for an extended period, the threshold for abnormality detection can be appropriately raised to reduce false positives. When possible abnormal signs are detected, the threshold can be lowered to increase sensitivity and catch anomalies in advance.

[0070] Weighted voting is an effective method for fusing the results of multiple classifiers. Based on the accuracy of the first, second, and third classifications, the three classifiers are assigned different weights, with the weights proportional to the classification accuracy. For each sample to be classified, the weighted voting results of each classifier are calculated, and the category with the highest number of votes is selected as the final classification result. When the votes for multiple categories are close, confidence analysis is used to calculate the difference between the highest and second highest voted categories. If the difference is less than a preset threshold, the category is marked as uncertain, triggering a more detailed analysis process.

[0071] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0072] (1) Convert the welding equipment status category into the subsystem status quantification value;

[0073] (2) For the abnormal temperature state, abnormal conveyor belt state, abnormal gas system state, and abnormal electrical system state in the welding equipment state category, assign subsystem weight coefficients to obtain weighted parameters of the subsystem state quantification value;

[0074] (3) Multiply the subsystem state quantification value by the subsystem weight coefficient to obtain the weighted state score;

[0075] (4) Extract the abnormal duration characteristics from the historical changes of welding equipment status categories and obtain the abnormal duration factor;

[0076] (5) Based on the correlation of multiple subsystem anomalies in the welding equipment status category, the degree of anomaly diffusion within the system is quantified to obtain the anomaly diffusion factor;

[0077] (6) The weighted status score, abnormal persistence factor and abnormal diffusion factor are comprehensively calculated to obtain the equipment health index. Multi-level warning thresholds are set according to the equipment health index to generate the welding equipment health assessment results.

[0078] Specifically, the state categories output by the state recognition model are numerically processed. Specifically, a normal operating state is assigned a value of 1.0, indicating a fully healthy equipment state; a mild abnormality is assigned a value of 0.8, indicating a minor abnormality but no impact on normal production; a moderate abnormality is assigned a value of 0.6, indicating a more severe abnormality requiring attention; and a severe abnormality is assigned a value of 0.3, indicating a severe abnormality requiring immediate intervention. State quantification values ​​are determined based on the impact of the reflow soldering equipment state on product quality and equipment safety. Values ​​closer to 1.0 indicate a state closer to normal operating conditions. This quantification method allows the states of different subsystems to be compared and comprehensively evaluated on a unified numerical scale. Different subsystem weight coefficients are assigned to the abnormal temperature, conveyor belt, gas system, and electrical system states within the soldering equipment state category. The subsystem weight coefficients reflect the impact of each subsystem on the overall health of the reflow soldering equipment. Weight assignment is based on two primary factors: the subsystem's impact on soldering quality and the severity of damage caused by the subsystem failure. The temperature system directly affects the solder melting and solidification process and plays a decisive role in soldering quality, so it is usually assigned a higher weight, such as 0.4. The conveyor system affects the stability and speed uniformity of the PCB during soldering and has a significant impact on soldering quality, so it is usually assigned the second highest weight, such as 0.3. The gas system mainly affects the oxidation level of the soldering environment and has a certain impact on soldering quality, so it is usually assigned a medium weight, such as 0.2. The electrical system, as the basis for equipment energy supply and control, is important but has a relatively small direct impact on soldering quality, so it is usually assigned a lower weight, such as 0.1. The sum of these weight coefficients must be equal to 1 to ensure the standardization of the scoring system.

[0079] Multiplying the subsystem status quantification value by the subsystem weight coefficient yields a weighted health score for each subsystem. The weighted health score is calculated as the subsystem status quantification value multiplied by the corresponding subsystem weight coefficient, reflecting the weighted contribution of each subsystem's status to the overall health status. For example, if the temperature system is moderately abnormal (quantified value 0.6), the conveyor system is slightly abnormal (quantified value 0.8), and both the gas system and the electrical system are normal (quantified value 1.0), then the weighted health score for the temperature system is 0.6 × 0.4 = 0.24, the weighted health score for the conveyor system is 0.8 × 0.3 = 0.24, the weighted health score for the gas system is 1.0 × 0.2 = 0.2, and the weighted health score for the electrical system is 1.0 × 0.1 = 0.1. The sum of the weighted health scores for each subsystem is 0.24 + 0.24 + 0.2 + 0.1 = 0.78, which serves as the baseline health score. Extracting the abnormal duration characteristics from the historical changes in the welding equipment status categories and obtaining the abnormal duration factor is an important part of health assessment. The abnormal duration factor reflects the persistence and evolution trend of the equipment's abnormal state. First, the cumulative percentage of time the equipment is in various abnormal states in the recent period (such as 24 hours) is counted, and then the abnormal duration factor is calculated based on the duration distribution of the abnormal state. The calculation of the abnormal duration factor considers two key parameters: one is the length of continuous abnormal time, and the other is the repetition frequency of the abnormal state. The longer the continuous time and the higher the repetition frequency, the more stable the abnormal state and the higher the risk of failure. The specific calculation method uses exponential weighted average, giving higher weights to abnormal events in the most recent time window and lower weights to abnormal events in the earlier time window, so as to more accurately reflect the severity and development trend of the current abnormality.

[0080] Based on the correlation of anomalies across multiple subsystems within the welding equipment status category, the anomaly diffusion factor (ADF) is quantified to determine the extent of anomaly propagation within the system. This DF describes whether a fault has spread from one subsystem to other subsystems, reflecting the systemic nature and chain reaction nature of the fault. The ADF is calculated based on a fault propagation network model between subsystems. First, a correlation matrix is ​​established between subsystems to represent the impact of a subsystem fault on other subsystems. The ADF is then calculated based on the current abnormal state of each subsystem and the ADF. For example, temperature anomalies often cause changes in conveyor belt tension, which in turn affects the dwell time of PCBs in various heating zones, thereby impacting temperature distribution. When multiple interconnected subsystems experience anomalies simultaneously, a higher ADF value indicates that the fault has spread within the system and the health condition is more critical.

[0081] The weighted status score, abnormal persistence factor, and abnormal diffusion factor are comprehensively calculated to obtain the equipment health index formula as follows:

[0082]

[0083] in, Indicates the device health index (Health Index), Represents the weight status value of the i-th subsystem, represents the weight factor of the ith subsystem, n represents the total number of subsystems, ADF represents the abnormal duration factor, and APF represents the abnormal propagation factor. 、 、 is the adjustment parameter, where The basic health score adjustment coefficient is usually set to 1.0; is the abnormal persistence influence coefficient, usually set to 0.2 to 0.5; is the abnormal diffusion influence coefficient, which is usually set to 0.3 to 0.6. In this formula, ( ) item to calculate the weighted status score of each subsystem, As the basic health score, ) represents the adjustment effect of abnormal persistence factor on health index, ( ) represents the adjustment effect of the abnormal diffusion factor on the health index. The product of the three comprehensively considers the combined impact of the current state, duration, and degree of diffusion on equipment health. Based on the calculated equipment health index, a multi-level warning threshold is set to generate a welding equipment health assessment result. The multi-level warning threshold is generally divided into three levels: a health index below 0.9 but above 0.7 is a mild warning, indicating that the equipment has a minor abnormality and requires attention; a health index below 0.7 but above 0.5 is a moderate warning, indicating that the equipment is significantly abnormal and requires inspection and maintenance; a health index below 0.5 is a severe warning, indicating that the equipment has a serious abnormality and requires immediate shutdown and maintenance. The warning threshold can be fine-tuned according to the specific equipment characteristics and production requirements to balance warning sensitivity and accuracy.

[0084] The health index calculation and early warning process is illustrated using a reflow soldering machine used to produce high-density PCBs. During a particular production run, the state recognition model detected temperature fluctuations in the reflow zone. The temperature gradient distribution map showed a temperature gradient of 7.8°C / cm in the central region, exceeding the normal value by 5°C / cm. This resulted in a moderate temperature anomaly, with a quantized value of 0.6. Simultaneously, the conveyor belt speed change rate increased, and the tension stability coefficient rose from the normal 0.05 to 0.09, indicating a mild conveyor belt anomaly with a quantized value of 0.8. The gas and electrical system parameters were normal, both with quantized values ​​of 1.0. Based on the preset subsystem weights (0.4 for the temperature system, 0.3 for the conveyor system, 0.2 for the gas system, and 0.1 for the electrical system), the weighted status scores for each subsystem were calculated as follows: 0.6 × 0.4 = 0.24 for the temperature system, 0.8 × 0.3 = 0.24 for the conveyor system, 1.0 × 0.2 = 0.2 for the gas system, and 1.0 × 0.1 = 0.1 for the electrical system, resulting in a basic health score of 0.78. A query of historical equipment data revealed three temperature anomalies within the past 12 hours, each lasting approximately 15 minutes. The calculated anomaly persistence factor was 0.25. The temperature anomaly and the conveyor anomaly occurred simultaneously, with a clear temporal correlation. The conveyor anomaly was typically detected within 5-10 minutes of the temperature anomaly, indicating that the anomaly had spread from the temperature system to the conveyor system. The calculated anomaly diffusion factor was 0.4. Substituting these factors into the health index calculation formula, with alpha = 1.0, beta = 0.3, and gamma = 0.4, we obtain a health index of 0.78 × (1 - 0.3 × 0.25) × (1 - 0.4 × 0.4) = 0.67, triggering a moderate alert. The resulting health assessment results include detailed information such as the current health index of 0.67, the alert level of "moderate alert," the abnormal subsystems "temperature system (moderate abnormality)" and "conveyor belt system (mild abnormality)," and the recommended measures: "Check the calibration of the reflow zone heating element and temperature sensor, and check the conveyor belt tension adjustment mechanism."

[0085] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0086] (1) Correlate the welding equipment health assessment results with the equipment parameters at historical time points to obtain a time series mapping database;

[0087] (2) Extract the fault evolution pattern from the time series mapping database, identify the characteristic change law before the equipment health index decreases, and obtain the fault precursor feature set;

[0088] (3) Perform nonlinear amplification processing on the weak change signal in the fault precursor feature set to obtain the amplified fault symptom feature;

[0089] (4) Collect typical evolution paths and characteristic sequences of different types of faults to form a reflow soldering characteristic timing pattern library;

[0090] (5) Calculate the similarity between the amplified fault symptom features and the reflow soldering feature timing pattern library to identify potential fault types and development trends, and obtain a fault matching score;

[0091] (6) Based on the fault matching score and the changing trend of the equipment health index, the equipment status and possible failure risks at future time points are predicted, and welding equipment failure warning information is generated.

[0092] Specifically, the welding equipment health assessment results are correlated with equipment parameters at historical time points to generate a time series mapping database. This process aligns and correlates assessment results, such as the health index and warning level, with the raw sensor data, feature matrix, and state classification results at the corresponding time points. For each time point, multidimensional data, including the equipment health index, quantified values ​​of each subsystem's state, temperature gradient distribution maps, conveyor belt tension fluctuation characteristics, and time-varying gas concentration curves, is recorded. The correlation method uses time window matching technology to establish a mapping relationship between the health assessment results and the equipment parameters within the previous time window (e.g., 5-30 minutes), forming a structured time series database containing timestamps, parameter values, and health status. Extracting fault evolution patterns from the time series mapping database is a key step in identifying fault precursors. By analyzing parameter change trends before a health index drop event, the characteristic changes during the equipment's transition from normal to abnormal state can be identified. Specific methods include trend analysis, change point detection, and pattern recognition. Trend analysis uses time series decomposition techniques to decompose each parameter sequence into trend terms, periodic terms, and residual terms, focusing on the slope of the trend terms. Change point detection uses the CUSUM (cumulative sum) algorithm or the Page-Hinkley detector to identify sudden changes in the parameter sequence. Pattern recognition uses the Dynamic Time Warping (DTW) algorithm to search the parameter sequence for segments similar to known fault patterns. Through these analyses, characteristic changes that occur before the health index decreases are extracted, such as gradually increasing temperature gradients, increasing frequency of conveyor belt tension fluctuations, and decreasing proportion of stable gas concentration intervals, forming a set of fault precursor features.

[0093] Nonlinear amplification of weak signal changes in fault precursor signatures is an important means of improving the sensitivity of early fault detection. Because signal changes in the early stages of a fault are often subtle and difficult to directly identify, signal amplification techniques are needed to enhance these subtle changes. Nonlinear amplification methods include exponential amplification, wavelet transform coefficient amplification, and response surface amplification. The exponential amplification function exponentially amplifies small deviations outside the normal range, making them more pronounced. The wavelet transform performs multi-scale decomposition of the signal, amplifies the wavelet coefficients representing abnormal characteristics, and then reconstructs the signal. Response surface amplification amplifies the coupling effects between parameters by modeling the nonlinear relationships between them. After amplification, subtle precursor signals such as temperature fluctuations, tension fluctuations, and gas concentration drift become more pronounced, facilitating subsequent identification and matching. Compiling typical evolution paths and characteristic sequences for different fault types to form a reflow soldering characteristic time-series pattern library is a crucial step in establishing a knowledge base for fault early warning. This pattern library contains characteristic sequence templates for common fault types, each of which contains parameter change sequences before, during, and after a fault occurs. Typical faults include heating element aging, abnormal conveyor belt tension, gas system leakage, and electrical control system failures. For each fault, we record the typical variation patterns of parameters such as temperature gradient distribution, belt tension fluctuation characteristics, and gas concentration time-varying curves during its occurrence, as well as the speed and severity of the fault's development. Each fault template in the pattern library contains information such as fault type, characteristic sequence, development speed, and severity, providing a reference standard for fault identification.

[0094] Computing the similarity between the amplified fault symptom signature and the reflow soldering feature timing pattern library is a key step in identifying potential fault types. This similarity calculation utilizes various distance metrics, including Euclidean distance, DTW distance, and Mahalanobis distance. Euclidean distance is suitable for comparing feature sequences of equal length; DTW distance can handle sequences of unequal length and with timeline distortion; and Mahalanobis distance considers inter-feature correlation and is suitable for comprehensive comparisons of multi-dimensional features. For each possible fault type, the similarity between the amplified fault symptom signature and the corresponding fault template in the pattern library is calculated to generate a fault matching score. The matching score ranges from 0 to 1, with higher values ​​indicating a higher match and a higher likelihood of that type of fault. Based on the fault matching score and the changing trend of the device health index, the device status and potential failure risk at future points in time are predicted. The prediction method integrates pattern matching and time series prediction techniques. First, the most likely fault type is determined based on the fault matching score. Then, the typical evolution path of that fault type is considered, combined with the historical changing trend of the device health index, to predict the future device status. The prediction results include the possible fault type, probability of failure, expected time of occurrence, and potential impact range. Finally, welding equipment failure warning information containing the above information is generated to provide guidance for equipment maintenance decisions.

[0095] For example, in the case of reflow soldering equipment at an electronics manufacturer, the equipment health monitoring system detected a gradual increase in the temperature fluctuation amplitude in the central reflow zone from the normal ±2°C to ±3.5°C, accompanied by a slight increase in the temperature gradient rate. These changes were recorded in a time-series mapping database. Analysis revealed that the heating element current exhibited small periodic fluctuations before the temperature fluctuations increased. Fault evolution pattern analysis identified this characteristic pattern: initial small current fluctuations, followed by increasing temperature fluctuations, and finally an abnormal temperature gradient and a decrease in the health index. Nonlinear amplification of the current fluctuation signal revealed a 2Hz pulsation with gradually increasing amplitude. Comparing this amplified feature with the pattern library revealed a 0.87 match with the "poor contact of the heating element" fault mode, indicating a high probability of occurrence. Based on the typical evolution path of this fault, it is predicted that without intervention, the equipment health index will drop from the current 0.85 to below 0.6 within the next 24 hours, generating a fault warning message: the risk of poor contact of the heating element, with a matching degree of 87%, is expected to develop into a moderate abnormality within 24 hours. It is recommended to check the contact condition of the heating element in the middle of the reflow zone after the end of today's production, so that maintenance personnel can take timely intervention measures before the fault seriously affects production.

[0096] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0097] (1) Map and associate welding equipment fault warning information with the equipment physical structure and operation logic to obtain equipment maintenance decision data;

[0098] (2) Analyze the correlation between historical welding quality data and equipment status parameters to obtain the quality impact factor matrix;

[0099] (3) Calculate equipment downtime losses, maintenance labor costs, and spare parts costs for different maintenance strategies to obtain a maintenance cost estimation table;

[0100] (4) Using the quality impact factor matrix and maintenance cost estimation table as constraints, the optimal maintenance time window is calculated through the integer programming algorithm to obtain the maintenance schedule;

[0101] (5) Generate specific maintenance items and operation procedures based on the fault type and equipment structure information in the welding equipment fault warning information, and obtain maintenance operation guidelines;

[0102] (6) Integrate the maintenance schedule with the maintenance operation guide to generate a welding equipment maintenance plan that includes maintenance time, maintenance items, required personnel, and resource allocation.

[0103] Specifically, welding equipment fault warning information is mapped and associated with the equipment's physical structure and operating logic. This step transforms abstract fault warning information into specific equipment components and functional units, establishing a mapping relationship between fault types and maintenance targets. First, based on the reflow soldering equipment's structural diagram and functional block diagram, the equipment is decomposed into key functional modules, such as the heating system, conveyor system, gas system, and electrical control system. Each module is then further decomposed into specific components and subassemblies. Through this hierarchical decomposition, a mapping table is established between fault types and specific maintenance targets. For example, a temperature abnormality fault might be mapped to the heating element, temperature sensor, or temperature controller; a conveyor abnormality fault might be mapped to the conveyor motor, tension adjustment device, or transmission gear. This mapping and association transforms fault warning information into clear maintenance decision-making data, including the specific components to inspect, possible causes of the problem, and recommended remediation methods. Analyzing the correlation between historical welding quality data and equipment status parameters is a key basis for optimizing maintenance strategies. By collecting welding quality inspection data (such as solder joint quality, weld strength, and void rate) and corresponding equipment status parameters (such as temperature distribution, conveyor belt tension, and gas concentration) over a period of time, we use statistical methods such as the Pearson correlation coefficient and the Spearman rank correlation coefficient to calculate the correlation between each parameter and quality indicators. Based on the analysis results, we create a quality impact factor matrix, in which each element represents the degree of influence of a certain equipment parameter on a specific quality indicator. This matrix helps identify which equipment components have the greatest impact on product quality, allowing these critical components to be prioritized when maintenance resources are limited.

[0104] Calculating the economic costs of different maintenance strategies is a key consideration in maintenance decision-making. Maintenance strategies include immediate maintenance, deferred maintenance, planned downtime maintenance, and in-service partial maintenance. For each strategy, three key costs are calculated: equipment downtime losses (including production delays, order deferrals, and lost capacity), maintenance labor costs (including labor hours, overtime, and professional technical support), and spare parts costs (including component prices, transportation costs, and storage fees). A cost model is established using historical maintenance records and financial data to estimate costs for different maintenance strategies. This table provides economic indicators for subsequent maintenance decisions. Using the quality impact factor matrix and the maintenance cost estimate table as constraints, an integer programming algorithm is used to calculate the optimal maintenance time window. The objective function of the integer programming model is to minimize total cost (the sum of maintenance cost and quality loss cost). Constraints include production planning requirements, resource availability, and maintenance time window constraints. By solving the integer programming problem, the lowest-cost maintenance schedule, including maintenance start and end times and temporal relationships, is determined while meeting quality requirements.

[0105] Generate specific maintenance items and operating procedures based on the fault type and equipment structure information in the welding equipment fault warning information. The maintenance items list in detail the specific parts that need to be inspected, replaced or adjusted, and the operating procedures describe the execution steps, required tools and safety precautions of the maintenance activities. The maintenance operation guide usually includes equipment function verification steps to ensure that the equipment can operate normally after maintenance and meet the design performance requirements. Integrate the maintenance schedule with the maintenance operation guide to generate a complete maintenance plan. The maintenance plan includes maintenance time (date and time period), maintenance items (specific parts and processing methods), required personnel (quantity, skill requirements) and resource allocation (tools, spare parts, equipment). The maintenance plan is presented in the form of a structured document to facilitate execution management and progress tracking.

[0106] For example, a reflow soldering machine's fault warning system detected a risk of poor contact between heating elements, with an 87% accuracy rate. This warning information was linked to the equipment structure map, identifying a specific group of heating elements and their electrical connections in the middle of the reflow zone for inspection. A quality impact analysis revealed a high correlation between temperature stability in this area and solder joint quality (correlation coefficient 0.82). Maintenance cost estimates indicated that immediate downtime for maintenance would be costly (including delays in the current batch), while scheduling maintenance after the end of today's production would reduce costs by 50%. Integer programming calculations determined the optimal maintenance window to be between 10:00 PM and 2:00 AM the following day after the end of that day's production. Maintenance instructions were generated based on the fault type, including cooling system disassembly steps, heating element inspection methods, and electrical contact treatment procedures. The resulting maintenance plan included a specific timeline, maintenance team configuration (one electrical engineer and two repair technicians), and a list of required resources (contact cleaner, thermal grease, testing equipment, etc.), ensuring efficient maintenance and eliminating fault risks before they impact production.

[0107] The above describes the operating status monitoring method of the reflow soldering equipment in the embodiment of the present application. The following describes the operating status monitoring system of the reflow soldering equipment in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the reflow soldering equipment operating status monitoring system includes:

[0108] An acquisition module is used to collect operating parameters of the reflow soldering equipment through multi-point sensors to obtain a soldering equipment status data set;

[0109] a correlation module for extracting a temperature gradient distribution map, conveyor belt tension fluctuation characteristics, and a gas concentration time-varying curve based on the welding equipment status data set, and performing a multi-dimensional feature cross-correlation analysis to obtain a welding equipment operation feature matrix;

[0110] A construction module is used to construct a state recognition model based on the welding equipment operation feature matrix, classify the equipment operation state through a dynamic threshold adaptive adjustment mechanism, and obtain a welding equipment state category;

[0111] A calculation module, configured to calculate an equipment health index according to the welding equipment status category, and set a multi-level warning threshold to obtain a welding equipment health assessment result;

[0112] A comparison module is used to construct a fault precursor feature extraction and amplification model using the welding equipment health assessment results, predict future states by comparing with the reflow soldering feature timing pattern library, and obtain welding equipment fault warning information;

[0113] A generation module is used to build a multi-objective maintenance strategy optimization system based on the welding equipment fault warning information, generate a maintenance plan through correlation analysis between welding quality and equipment status, and obtain a welding equipment maintenance plan.

[0114] Through the collaborative cooperation of the above components, a comprehensive welding equipment status data set is obtained by collecting operating parameters through multi-point sensors, realizing comprehensive monitoring of multiple systems and multiple parameters of the equipment. Compared with the traditional single-point monitoring method, the comprehensiveness and accuracy of data collection are greatly improved, providing a rich data basis for subsequent analysis; the temperature gradient distribution map, conveyor belt tension fluctuation characteristics and gas concentration time-varying curve are extracted, and multi-dimensional feature cross-correlation analysis is performed to obtain the welding equipment operation feature matrix. This process fully explores the inherent correlation between multi-dimensional parameters and reveals the mutual influence mechanism between different subsystems, so that abnormal state detection is no longer limited to the judgment of a single parameter exceeding the limit, but can be evaluated from the overall performance of the system; the state recognition model constructed based on the welding equipment operation feature matrix realizes accurate classification of the equipment operation status through a dynamic threshold adaptive adjustment mechanism. This mechanism dynamically adjusts the classification threshold according to the state transition frequency and duration in the equipment operation history data, effectively solving the misjudgment problem caused by fixed thresholds, and improving the accuracy and stability of state recognition. In particular, in the combined application of the three machine learning algorithms of support vector machine, random forest and gradient boosting decision tree, the respective advantages of different algorithms are fully utilized. The support vector machine is good at processing high The random forest algorithm has good noise resistance and feature importance assessment capabilities, while the gradient boosting decision tree is good at processing unbalanced data and capturing subtle feature changes. The weighted fusion of the three algorithms significantly improves classification performance. The equipment health index is calculated according to the welding equipment status category, and multi-level warning thresholds are set, realizing the transition from status classification to health quantification, making equipment status assessment more precise and intuitive. The fault precursor feature extraction and amplification model is constructed using the welding equipment health assessment results. The future status is predicted by comparing the reflow soldering feature time series pattern library, transforming passive response into active prediction, greatly improving the fault warning time and gaining sufficient preparation time for maintenance decision-making. A multi-objective maintenance strategy optimization system is constructed based on welding equipment fault warning information. Maintenance plans are generated by analyzing the correlation between welding quality and equipment status. For the first time, product quality factors are introduced into the maintenance decision-making process, realizing the transition from purely equipment-oriented to coordinated optimization of quality and cost. The system solves the optimal maintenance time window through an integer programming algorithm, minimizing the total maintenance cost while ensuring welding quality. This provides scientific and economical maintenance decision support for enterprises, significantly improving production efficiency and product yield, and reducing maintenance costs and unplanned downtime losses.

[0115] Reference Figure 3In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, memory, display screen, input device, network interface and database connected via a system bus. The processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the above method is implemented.

[0116] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0117] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-described method when executed by a processor. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0118] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware using a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media provided herein and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM.

[0119] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0120] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0121] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for monitoring the operating status of a reflow soldering device, characterized in that: include: The operating parameters of the reflow soldering equipment are collected by multi-point sensors to obtain the soldering equipment status data set; Based on the welding equipment status data set, the temperature gradient distribution map, conveyor belt tension fluctuation characteristics and gas concentration time-varying curve are extracted, and multi-dimensional feature cross-correlation analysis is performed to obtain the welding equipment operation feature matrix; A state recognition model is constructed based on the welding equipment operation feature matrix, and the equipment operation status is classified through the dynamic threshold adaptive adjustment mechanism to obtain the welding equipment state category; including five typical operation states of predefined reflow soldering equipment: normal operation state, temperature abnormal state, conveyor belt abnormal state, gas system abnormal state and electrical system abnormal state; the welding equipment operation feature matrix is ​​divided into a training matrix and a verification matrix; the training matrix is ​​processed by the support vector machine algorithm to construct the first classifier, and the performance of the first classifier is evaluated with the verification matrix to obtain the first classification accuracy; the training matrix is ​​processed by the random forest algorithm to construct the second The first classifier is a two-classifier, and the performance of the second classifier is evaluated with a validation matrix to obtain the second classification accuracy; the training matrix is ​​processed by a gradient boosting decision tree algorithm to construct a third classifier, and the performance of the third classifier is evaluated with a validation matrix to obtain the third classification accuracy; a dynamic threshold adaptive adjustment mechanism is constructed to dynamically adjust the discrimination threshold parameters of the first classifier, the second classifier, and the third classifier according to the state transition frequency and duration in the equipment operation history data; based on the first classification accuracy, the second classification accuracy, and the third classification accuracy, the discrimination results of the three classifiers are fused by a weighted voting method to obtain the welding equipment state category; Calculate the equipment health index according to the welding equipment status category, and set a multi-level warning threshold to obtain the welding equipment health assessment result, including: converting the welding equipment status category into a subsystem status quantification value; assigning a subsystem weight coefficient to the temperature abnormality, conveyor belt abnormality, gas system abnormality and electrical system abnormality in the welding equipment status category to obtain a weighted parameter of the subsystem status quantification value; multiplying the subsystem status quantification value by the subsystem weight coefficient to obtain a weighted status score; extracting the abnormality duration feature from the historical changes of the welding equipment status category to obtain an abnormality duration factor; based on the correlation of multiple subsystem abnormalities in the welding equipment status category, quantify the degree of abnormality diffusion in the system to obtain an abnormality diffusion factor; comprehensively calculate the weighted status score, abnormality duration factor and abnormality diffusion factor to obtain the equipment health index, and set a multi-level warning threshold according to the equipment health index to generate a welding equipment health assessment result; The results of welding equipment health assessment are used to build a fault precursor feature extraction and amplification model. The future state is predicted by comparing the reflow soldering feature timing pattern library to obtain welding equipment fault warning information. A multi-objective maintenance strategy optimization system is constructed based on welding equipment fault warning information. A maintenance plan is generated by analyzing the correlation between welding quality and equipment status to obtain a welding equipment maintenance plan.

2. The method for monitoring the operating status of reflow soldering equipment according to claim 1, wherein: Multi-point sensors on the reflow soldering equipment collect operating parameters to generate a soldering equipment status dataset. This includes installing temperature sensor arrays in the preheating, reflow, and cooling zones of the reflow soldering equipment, speed sensors and tension sensors in the conveyor belt system, gas concentration sensors and air pressure sensors in the gas supply system, and current and voltage sensors in the power supply system, forming a multi-type sensor network. An industrial Internet of Things (IIoT) module is used to uniformly sample the data collected by the multi-type sensor network. The sampling frequencies for temperature data, conveyor belt parameters, gas parameters, and electrical parameters are 5Hz, 2Hz, and 20Hz, respectively. The collected temperature data is denoised using a sliding average filter algorithm, the conveyor belt speed data is denoised using a median filter algorithm, and the gas concentration data is processed using a Kalman filter algorithm. Missing data points are supplemented using linear interpolation or polynomial interpolation based on their consecutive missing lengths, and outliers are preliminarily screened using the statistical 3σ principle. The processed sensor data are aligned according to a unified timestamp to generate a soldering equipment status dataset.

3. The method for monitoring the operating status of reflow soldering equipment according to claim 1, wherein: The temperature gradient distribution map, conveyor belt tension fluctuation characteristics and gas concentration time-varying curve are extracted from the welding equipment status data set, and a multi-dimensional feature cross-correlation analysis is performed to obtain the welding equipment operation feature matrix, including: segmenting the welding equipment status data set according to different time window lengths, and the window length is adaptively adjusted to 5 minutes to 30 minutes according to different equipment operation cycles; calculating the temperature gradient distribution map for the temperature data in each time window, and obtaining the spatial temperature gradient by dividing the difference between adjacent temperature sensors by the sensor spacing, and then combining the time dimension to form the temperature gradient spatiotemporal distribution matrix; extracting the conveyor belt tension fluctuation characteristics from the conveyor belt parameters. The belt tension fluctuation characteristics, including the peak-to-peak tension, tension stability coefficient and tension pulsation frequency characteristics, are analyzed, and the tension fluctuation feature vector is constructed; the gas concentration data is transformed into a time-frequency form to extract the time-varying curve characteristics of the gas concentration, including the gas concentration change rate, the gas concentration fluctuation period and the gas concentration stable interval ratio; the phase correlation between the temperature gradient distribution map and the conveyor belt tension fluctuation characteristics is calculated, the time-lag correlation between the gas concentration time-varying curve and the temperature gradient is analyzed, and the parameter cross-influence matrix is ​​constructed; the extracted features are reduced in dimensionality by the principal component analysis method, and the principal components with a cumulative contribution rate of 95% are retained to form the welding equipment operation characteristic matrix.

4. The method for monitoring the operating status of reflow soldering equipment according to claim 1, wherein: The health assessment results of welding equipment are used to construct a fault precursor feature extraction and amplification model, and the future state is predicted by comparing with the reflow soldering feature timing pattern library to obtain welding equipment fault warning information, including: correlating the welding equipment health assessment results with the equipment parameters at historical time points to obtain a timing mapping database; extracting the fault evolution pattern from the timing mapping database, identifying the characteristic change law before the equipment health index decreases, and obtaining a fault precursor feature set; performing nonlinear amplification processing on the weak change signal in the fault precursor feature set to obtain the amplified fault symptom feature; collecting the typical evolution paths and feature sequences of different types of faults to form a reflow soldering feature timing pattern library; calculating the similarity between the amplified fault symptom feature and the reflow soldering feature timing pattern library, identifying the potential fault type and development trend, and obtaining a fault matching score; based on the fault matching score and the equipment health index change trend, predicting the equipment state and possible failure risk at future time points, and generating welding equipment fault warning information.

5. The method for monitoring the operating status of reflow soldering equipment according to claim 1, wherein: A multi-objective maintenance strategy optimization system is constructed based on welding equipment fault warning information. A maintenance plan is generated through correlation analysis between welding quality and equipment status to obtain a welding equipment maintenance plan, including: mapping and correlating welding equipment fault warning information with the equipment physical structure and operation logic to obtain equipment maintenance decision data; analyzing the correlation between historical welding quality data and equipment status parameters to obtain a quality impact factor matrix; calculating equipment downtime losses, maintenance labor costs and spare parts costs for different maintenance strategies to obtain a maintenance cost estimation table; using the quality impact factor matrix and the maintenance cost estimation table as constraints, the optimal maintenance time window is calculated through an integer programming algorithm to obtain a maintenance schedule; generating specific maintenance items and operation procedures based on the fault type and equipment structure information in the welding equipment fault warning information to obtain a maintenance operation guide; integrating the maintenance schedule with the maintenance operation guide to generate a welding equipment maintenance plan that includes maintenance time, maintenance items, required personnel and resource allocation.

6. A reflow soldering equipment operating status monitoring system, used to implement the reflow soldering equipment operating status monitoring method according to any one of claims 1 to 5, characterized in that: The reflow soldering equipment operation status monitoring system includes: An acquisition module is used to collect operating parameters of the reflow soldering equipment through multi-point sensors to obtain a soldering equipment status data set; The correlation module is used to extract the temperature gradient distribution map, conveyor belt tension fluctuation characteristics and gas concentration time-varying curve based on the welding equipment status data set, and perform multi-dimensional feature cross-correlation analysis to obtain the welding equipment operation feature matrix; A construction module is used to build a state recognition model based on the welding equipment operation feature matrix, classify the equipment operation state through a dynamic threshold adaptive adjustment mechanism, and obtain the welding equipment state category; The calculation module is used to calculate the equipment health index according to the welding equipment status category and set multi-level warning thresholds to obtain the welding equipment health assessment results; The comparison module is used to build a fault precursor feature extraction and amplification model based on the welding equipment health assessment results. It predicts the future state by comparing the reflow soldering feature timing pattern library to obtain welding equipment fault warning information; The generation module is used to build a multi-objective maintenance strategy optimization system based on welding equipment fault warning information, generate a maintenance plan through correlation analysis between welding quality and equipment status, and obtain a welding equipment maintenance plan.

7. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and is characterized in that when the processor executes the computer program, the method for monitoring the operating status of the reflow soldering equipment according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the processor is caused to execute the method for monitoring the operating status of reflow soldering equipment according to any one of claims 1 to 5.

Images