SMT production line analysis method and system based on real-time data acquisition

Abstract

The present application relates to the technical field of state monitoring, in particular to a SMT production line analysis method and system based on real-time data acquisition, comprising the following steps: calling a multi-device operation log to determine a device connection relationship, analyzing a track gauge and pressure change marker structure turning point, identifying a rate offset temperature control path code generation direction sequence, analyzing an abnormal signal sequence to generate a path identifier, matching a migration direction according to a state score change to predict a fault trend, and generating a fault trend prediction result, wherein, in the present application, a device connection relationship is constructed based on the time sequence and interval of the operation log, a process link is dynamically determined, a direction combination marker structure turning point is determined based on the track gauge and pressure, a structure section is packaged, a dominant path behavior is determined based on the direction combination and repetition of the rate offset temperature control, a main cause chain starting point is positioned based on the abnormal signal starting point and coverage section, and a fault evolution prediction is output based on the score sequence and migration direction matching.

Description

Technical Field

[0001] This invention relates to the field of condition monitoring technology, and in particular to an SMT production line analysis method and system based on real-time data acquisition. Background Technology

[0002] The field of condition monitoring technology encompasses the process of collecting, analyzing, and diagnosing the operating status of industrial equipment. Its core content lies in collecting operational data through front-end acquisition devices such as sensors, processing information using data acquisition systems, and achieving comprehensive monitoring of equipment operating conditions, operating parameters, and abnormal behaviors. It covers key components such as hardware sensing layer, data transmission link, data analysis platform, and human-machine interface, aiming to achieve full-process perception and status cognition of industrial production systems through real-time data acquisition and analysis, and is applied to process management and equipment management in multiple fields such as manufacturing, energy, and transportation.

[0003] Among them, the SMT production line analysis method based on real-time data acquisition refers to a control method that collects and analyzes status data of multiple key nodes in the surface mount production line. It acquires real-time information on the working status, cycle time, material supply, and product yield of equipment such as pick-and-place machines, reflow soldering machines, and printers in the production line. By constructing data acquisition units, time synchronization mechanisms, and node identification systems, and relying on unified communication protocols and edge computing units, the data generated by each device is aggregated and structured. The collected data is sorted, mapped, and classified using a rule analysis engine, and the production line status is periodically analyzed in conjunction with the production cycle time model. The method usually integrates and parses multi-dimensional information through data acquisition cards and industrial control terminals deployed at the end of the production line or in key links, forming an information infrastructure suitable for production line management.

[0004] Traditional SMT production line analysis technology mainly relies on static classification and rule analysis of sensor-collected data. The connection status of production line equipment is usually fixedly defined by manually marked process templates or communication protocol templates, lacking dynamic identification of operating sequence and response timing. It cannot adapt quickly to scenarios of adding or removing equipment or restructuring the production line. Furthermore, the structural division of production batches mainly relies on time segmentation or manual recording. When key process parameter adjustments are not recorded synchronously, it is impossible to accurately identify the boundaries of data segments. During the response to concurrent abnormal signals, it is impossible to determine the master-slave relationship between signals, and only provides prompts through single-point alarms or time sequence. There is a lack of sorting out the correlation chain between abnormalities, resulting in fault scores being unrelated to path changes and making it difficult to form multi-cycle prediction results with the ability to identify evolutionary trends. Summary of the Invention

[0005] To address the technical problems existing in the prior art, embodiments of the present invention provide an SMT production line analysis method based on real-time data acquisition, comprising the following steps:

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an SMT production line analysis method based on real-time data acquisition, comprising the following steps:

[0007] S1: Call the multi-device operation log, determine the time sequence of the completion signal and start signal of multiple devices, calculate the time interval between the signals and compare it with the communication beat standard, identify the connection relationship between devices based on the time sequence and time interval, and generate device connection structure information;

[0008] S2: Using the device connection structure information to identify the mounting process nodes, call and analyze the direction of change of the gauge and pressure parameters of the corresponding nodes in adjacent cycles, identify parameter adjustment behavior, mark it as a data structure turning point, divide the data into multiple structural segments, and generate parameter structure encapsulation results.

[0009] S3: Call the parameter structure encapsulation result, compare the change direction of mounting rate, visual offset, and temperature control feedback, identify the combined path code with the same direction, analyze the repetition frequency, determine the behavior characteristics of the dominant running path, and generate a directional path combination sequence.

[0010] S4: Using the aforementioned direction path combination sequence, analyze the recording time of multiple abnormal signals, compare the signal triggering order and duration segment, determine the response cycle coverage relationship, and record the direction combination in accordance with the dominant path behavior characteristic conditions to generate abnormal path signal identifiers.

[0011] As a further aspect of the present invention, the device connection structure information includes node connection order, time interval between nodes, and connection path direction; the parameter structure encapsulation result includes structure segment number, encapsulation start time, and parameter change direction combination; the direction path combination sequence specifically includes direction encoding type, path combination number, and repetition frequency within the period; and the abnormal path signal identifier specifically refers to the abnormal signal start order, coverage period range, and main attribution signal type.

[0012] As a further aspect of the present invention, the communication timing standard refers to the signal response interval specification that should be met between multiple devices during collaborative operation, reflecting the standard time configuration between the device completion signal and the lower-level device start signal;

[0013] The dominant path behavior characteristic condition refers to the high-frequency combination trend of three state characteristics, namely, mounting rate, visual offset and temperature control feedback, within the operating cycle. By statistically analyzing the repetition frequency of the direction code and identifying its duration, the direction path combination with a repetition frequency higher than a preset frequency threshold and a duration exceeding a preset period threshold is defined as the dominant path.

[0014] As a further aspect of the present invention, the step of obtaining the device connection structure information is as follows:

[0015] S101: Call the multi-device operation log, collect the start signal and completion signal time information of each device in the SMT production line, including pick and place machine, reflow soldering machine, and printer, sort and match the time information, analyze the time series of device operation, and generate the device signal sequence sequence;

[0016] S102: Based on the sequence of device signals, calculate the time interval between the start signal and the completion signal between adjacent devices, compare the time interval with the communication beat standard, establish the connection path between nodes, and obtain the connection relationship coefficient between nodes;

[0017] S103: Based on the connection relationship coefficients between the nodes, summarize the connection order, node-corresponding time interval and connection direction between each device node, construct the operation process topology structure, and obtain device connection structure information.

[0018] As a further aspect of the present invention, the step of obtaining the parameter structure encapsulation result is as follows:

[0019] S201: Using the equipment connection structure information to identify the mounting process node, call and analyze the gauge and pressure parameters of the corresponding node, calculate the change status in each production cycle, perform direction judgment on the change value of gauge parameter and pressure parameter in each cycle, filter parameter adjustment behaviors with the same change direction in the same cycle, and obtain parameter adjustment combination sequence.

[0020] S202: Based on the parameter adjustment combination sequence, analyze the distribution of parameter change trends in adjacent production cycles, determine the relationship between each set of parameter adjustment behaviors and the corresponding production cycle, mark the positions where parameter adjustments occur within the cycle as data structure turning points, and obtain a structural turning point sequence.

[0021] S203: Call the structural turning point sequence, divide the production line data into multiple structural segments with the turning point as the boundary, encapsulate the data according to the structural segment, establish structural segment data with structural segment number, encapsulation start time, and parameter direction combination code, and obtain the parameter structure encapsulation result.

[0022] As a further aspect of the present invention, the step of obtaining the direction path combination sequence is as follows:

[0023] S301: Call the parameter structure encapsulation result, determine the direction of change of the mounting rate, visual offset and temperature control feedback in each structural segment, convert the change trend of each parameter into a direction code, and generate a parameter direction combination code by combining the three direction codes in each cycle.

[0024] S302: Based on the parameter direction combination coding, identify the combination path coding with consistent parameter change direction characteristics, count the frequency of each direction combination coding in the corresponding structural segment, analyze the occurrence trend of each combination, and obtain the path coding repetition frequency coefficient.

[0025] S303: Based on the path code repetition frequency coefficient, and in accordance with the behavioral characteristic conditions of the dominant operating state path, filter the directional combination codes, mark the codes that meet the conditions as structural identifiers, organize them into path sequence numbers, establish the dominant behavioral combination of the periodic path, and obtain the directional path combination sequence.

[0026] As a further aspect of the present invention, the step of obtaining the abnormal path signal identifier is as follows:

[0027] S401: Based on the direction path combination sequence, filter the recording time of abnormal signals in the mounting rate, feeding signal, and temperature control signal, analyze the triggering time sequence of the signals within the same period, cross-compare the start time and duration of each group of signals, and obtain abnormal signal time sequence distribution data.

[0028] S402: Based on the abnormal signal time-series distribution data, determine the triggering order between each signal, perform overlapping interval detection on the start time and duration of each group of signals, filter out signals that cover the response period of the other signals in the time series, and obtain the signal type code of the coverage interval;

[0029] S403: Based on the signal type encoding of the coverage area, mark the main signal type selected in each group of periodic signals and use it as the starting signal of the main attribution chain. Organize the number, type, and periodic segment of the main signal, establish the type sequence of the main attribution chain, and obtain the abnormal path signal identifier.

[0030] As a further aspect of the present invention, the method further includes:

[0031] S5: Based on the abnormal path signal identifier, analyze the placement rate fluctuation, material supply stabilization time and temperature control response time, calculate the cycle status score, analyze the relationship between score changes and abnormal type attribution, match the migration path direction, predict the equipment failure evolution path, and generate failure trend prediction results.

[0032] The specific fault trend prediction results include the score change segment identifier, the abnormal migration path number, and the fault type prediction result.

[0033] As a further aspect of the present invention, the step of obtaining the fault trend prediction result is as follows:

[0034] S501: Based on the abnormal path signal identifier, calculate the equipment status score for each cycle by analyzing the placement rate fluctuation amplitude, material supply stabilization time and temperature control response duration, and obtain the cycle status score sequence.

[0035] S502: Based on the cycle state scoring sequence, analyze the step-by-step change process of the continuous scoring results in each operating cycle, and statistically analyze the anomaly type attribution in each segment, identify the anomaly type path direction corresponding to each segment, and obtain the scoring change path distribution data.

[0036] S503: Based on the score change path distribution data, determine the migration path direction corresponding to the anomaly type in multiple score change segments, identify the migration path code and fault type number under each operating cycle, predict the fault evolution path of the equipment, and generate fault trend prediction results.

[0037] The SMT production line analysis system based on real-time data acquisition includes:

[0038] The node connection identification module calls the multi-device operation log to determine the time sequence of the completion and start signals of multiple devices, calculates the time interval between the signals and compares it with the communication beat standard, and identifies the connection relationship between devices based on the time sequence and time interval, generating device connection structure information.

[0039] The structural data segmentation module uses the equipment connection structure information to analyze the direction of change of gauge and pressure parameters in adjacent cycles during the mounting process, identify parameter adjustment behavior, and mark it as a data structure turning point, dividing the data into multiple structural segments and generating parameter structure encapsulation results.

[0040] The path behavior encoding module calls the parameter structure encapsulation result, compares the change direction of mounting rate, visual offset, and temperature control feedback, identifies the combined path codes with the same direction, analyzes the repetition frequency, judges the dominant running path behavior characteristics, and generates a directional path combination sequence.

[0041] The dominant path identification module uses the directional path combination sequence to analyze the recording time of multiple abnormal signals, compare the signal triggering order and duration, determine the response cycle coverage relationship, and generate abnormal path signal identifiers by comparing the recorded directional combinations with the dominant path behavior characteristic conditions.

[0042] The behavior trend prediction module analyzes the placement rate fluctuation, material supply stabilization time and temperature control response time based on the abnormal path signal identifier, calculates the cycle status score, analyzes the relationship between score changes and abnormal type attribution, matches the migration path direction, predicts the equipment failure evolution path, and generates failure trend prediction results.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0044] In this invention, the actual connection relationship between devices is dynamically established by combining and judging the time sequence and interval between the completion signal and the start signal of multiple devices, thus solving the problem of path recognition lag caused by changes in node structure. The structural turning point is identified by the directional change of the combination of gauge and pressure parameters. Production data is packaged in segments to enhance the local analytical capability of parameter behavior. By statistically analyzing the frequency of directional combinations of state characteristics such as mounting rate, visual offset, and temperature control feedback within a cycle, the dominant operating path is determined and its directional sequence is extracted. Based on the response overlap relationship between the trigger time and the continuous segment of the abnormal signal, the starting point of the main attribution chain is determined and the abnormal path identifier is associated. A periodic scoring sequence is established by combining mounting rate fluctuation, material supply response time and temperature control duration. The migration path direction is constructed by matching the score change with the abnormal type, thereby realizing the trend prediction of equipment failure and improving the path orientation of the evolution stage. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of the steps of the present invention;

[0047] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0048] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0049] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0050] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0051] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0052] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0053] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0054] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0055] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0056] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0057] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0058] Please see Figure 1 This invention provides an SMT production line analysis method based on real-time data acquisition, comprising the following steps:

[0059] S1: Call the multi-device operation log, determine the time sequence of the completion signal and start signal of multiple devices, calculate the time interval between the signals and compare it with the communication beat standard, identify the connection relationship between devices based on the time sequence and time interval, and generate device connection structure information;

[0060] S2: Using the equipment connection structure information to identify the mounting process nodes, call and analyze the change direction of the gauge and pressure parameters of the corresponding nodes in adjacent cycles, identify parameter adjustment behavior, mark it as the data structure turning point, divide the data into multiple structural segments, and generate parameter structure encapsulation results.

[0061] S3: Call the parameter structure encapsulation result, compare the change direction of mounting rate, visual offset, and temperature control feedback, identify the combined path code with the same direction, analyze the repetition frequency, determine the behavior characteristics of the dominant running path, and generate the directional path combination sequence.

[0062] S4: Utilize the directional path combination sequence to analyze the recording time of multiple abnormal signals, compare the signal triggering order and duration segment, determine the response cycle coverage relationship, and record the directional combination in accordance with the dominant path behavior characteristic conditions to generate abnormal path signal identifiers.

[0063] S5: Based on the abnormal path signal identifier, analyze the placement rate fluctuation, material supply stabilization time and temperature control response time, calculate the cycle status score, analyze the relationship between score changes and abnormal type attribution, match the migration path direction, predict the equipment failure evolution path, and generate failure trend prediction results.

[0064] The equipment connection structure information includes the node connection sequence, the time interval between nodes, and the connection path direction. The parameter structure encapsulation result includes the structure segment number, the encapsulation start time, and the parameter change direction combination. The direction path combination sequence specifically includes the direction encoding type, the path combination number, and the repetition frequency within the period. The abnormal path signal identifier specifically refers to the abnormal signal start sequence, the coverage period range, and the main attribution signal type. The fault trend prediction result specifically includes the score change segment identifier, the abnormal migration path number, and the fault type prediction result.

[0065] Communication timing standard refers to the signal response interval specification that should be met between multiple devices during collaborative operation, reflecting the standard time configuration between the completion signal of a device and the start signal of the next-level device;

[0066] The dominant path behavior characteristic condition refers to the high-frequency combination trend of three state characteristics, namely placement rate, visual offset and temperature control feedback, within the operating cycle. By statistically analyzing the repetition frequency of direction codes and identifying their duration, the direction path combination with a repetition frequency higher than a preset frequency threshold and a duration exceeding a preset cycle threshold is defined as the dominant path.

[0067] Please see Figure 2 The steps for obtaining device connection structure information are as follows:

[0068] S101: Call the multi-device operation log, collect the start signal and completion signal time information of each device in the SMT production line, including pick and place machine, reflow soldering machine, and printer, sort and match the time information, analyze the time series of device operation, and generate the device signal sequence sequence;

[0069] Accessing multi-device operation logs refers to reading the historical operation data files recorded by the control systems of typical equipment in an SMT production line, such as pick-and-place machines, reflow soldering machines, and printers. This involves extracting the timestamp information corresponding to the start and finish signals for each device. For example, pick-and-place machine A's start time is 08:00:00 and finish time is 08:01:45, while reflow soldering machine B's start time is 08:02:00 and finish time is 08:03:30. By obtaining the operation logs of each device and parsing the time field of each event, all events are unified into a timeline. This timeline is then sorted in ascending order by time field to form a complete time sequence. Finally, start and finish events with similar times are paired for each device to ensure a one-to-one correspondence. For example, if the start time of device A is... The completion time is Then mark its running cycle as This process is used to identify the operating segments of other devices in sequence, followed by analyzing the sequential relationships between each device record, such as device A's... At 08:01:45, device B The time is 08:02:00, indicating that device B's startup time is slightly later than device A's completion time. The order is A→B. If there is any overlap or intersection, the priority is determined by the order of signal times, for example, the startup time of device C. If the time is 08:01:40, which is earlier than that of device B, then its order is A→C→B. The above pairing and sorting operations are performed on each device. By constructing an ordered list of all running records, the final device signal sequence is generated.

[0070] S102: Based on the sequence of device signals, calculate the time interval between the start signal and the completion signal between adjacent devices, compare the time interval with the communication beat standard, establish the connection path between nodes, and obtain the connection relationship coefficient between nodes;

[0071] The specific formula for comparing the time interval with the communication cycle standard is as follows:

[0072] ;

[0073] Calculate the node connection difference value, compare the time interval with the communication clock standard, and establish the connection path between nodes;

[0074] in, For the first The device and the first The difference in node connectivity between individual devices. For the first The normalized value of the device start signal time. For the first The normalized value of the signal completion time of each device. For the first The device and the first The priority weight of each device connection relationship. For the first Normalized value of signal interval within a production cycle For the first The device and the first The reference communication clock standard normalized value for each device For reference period quantity, For the first The device and the first The normalized mean value of the standard time difference series of each device. This is the index of the previous device's number. For the number index of the next device, Index the production cycle number.

[0075] The specific formula for comparing the time interval with the communication cycle standard in the above content is as follows: Specifically, the time quantities within the same observation window are first normalized, and then the operations of subtraction, multiplication by weight, summation and division by the number of samples, subtraction with the benchmark, and taking the absolute value are performed in sequence. The operation logic is to normalize the difference between the trigger intervals of the current nodes. Dimensionless weight After adjustment, the interval deviation from each period within the window By summation and by sample size The average value is then combined and finally compared with the reference standard mean. The non-negative difference index is obtained by differentiating and taking the absolute value. The subtraction is used to obtain the timing difference, multiplication is used to inject the influence of path priority, summation is used to calculate the deviation within the aggregation cycle, division is used to obtain the mean magnitude, and the absolute value is used to eliminate the influence of sign direction and provide a scalar measure of the deviation intensity. To support the numerical source of each parameter, adjacent equipment on the same production line is selected. and For on-site monitoring data within a continuous production window, with intervals in seconds, the raw intervals are first collected, and then linear normalization is performed between the minimum and maximum intervals within the window. Obtain the dimensionless value and set The initial intervals are obtained from the difference between the completion and startup times in the device logs. For example, the interval for cycle 1 is 12 seconds, cycle 2 is 14 seconds, cycle 3 is 9 seconds, cycle 4 is 16 seconds, and cycle 5 is 20 seconds. The minimum interval within the observation window is... seconds, maximum interval Seconds, referencing the communication timer standard of 12 seconds, corresponding to The start and finish times at the current moment are normalized within the same window and then taken. , The path weight is matched to the normalized difference of approximately 0.33 corresponding to the currently observed 14-second difference. The dimensionless value is obtained by weighting the SMEMA handshake success rate and board transfer occupancy rate in this window. For example, if the handshake is successful 4 times instead of 5 times and the transfer occupancy rate is approximately 0.75, the result is obtained after weighted mapping. , The normalized mean value of the beat standard within this window is used as a reference. Since the standard is fixed, it is taken as... .

[0076] Table 1 Monitoring Window Interval and Normalization Table

[0077]

[0078] As shown in Table 1, the sum of the normalized differences over the five periods is... Substitute into the formula to calculate :

[0079] First, calculate the weighted current difference term:

[0080] ;

[0081] Adding this to the summation term, we get:

[0082] ;

[0083] Divide by get ,and work Take the absolute value .

[0084] The result is compared with the dimensionless baseline interval used for connectivity determination in this section. The baseline interval is set based on the empirical range that the normal operating window statistic of the same line in the past shift is not less than 0 and not greater than 0.20. If the node connection difference is within this interval, the result indicates that the node connection difference within this window is within the usable range. Combined with the connectivity determination strategy of this step, along with the subsequent step S103, this result can be used as the edge weight input to construct the running process topology and output the device connection structure information.

[0085] Formula passed and The joint metric combines the current time difference with the stability deviation within the window for correction, and uses... By performing benchmark alignment, comparable connectivity difference indices can be generated within the same dimensionless framework. .

[0086] S103: Based on the connection relationship coefficients between nodes, summarize the connection sequence, node-corresponding time interval and connection direction between each device node, construct the operation process topology, and obtain device connection structure information;

[0087] Based on the connection coefficients between nodes For each set of device connections, the connection order, time interval, and direction are extracted. For example, in the connection from A to B, the connection direction is recorded as positive and the time interval is 15 seconds. All device connection pairs are sorted in chronological order to form path chains, such as A→B→C→D, and the directionality of each path is labeled. A directed graph structure is constructed, with nodes as vertices and connecting paths as directed edges, with edge weights of 1 / 2. Value, if a certain path If the threshold is exceeded, a higher edge weight is assigned, forming a topology graph with time constraints. The in-degree and out-degree relationships of each node are further recorded, and the presence of unclosed paths or isolated nodes is screened. If device D has no connection to other devices, that node is excluded from the graph. Finally, the time interval, directionality, and signal path number of each node in the device connection chain are statistically analyzed and integrated into a topology matrix. ,in Indicates from node To the node By analyzing the connection relationships and direction indicators, the device connection structure information can be obtained.

[0088] Please see Figure 3 The steps to obtain the parameter structure encapsulation result are as follows:

[0089] S201: Using the equipment connection structure information to identify the mounting process nodes, call and analyze the gauge and pressure parameters of the corresponding nodes, calculate the change status in each production cycle, perform direction judgment on the change values ​​of gauge parameters and pressure parameters in each cycle, filter parameter adjustment behaviors with the same change direction in the same cycle, and obtain parameter adjustment combination sequence.

[0090] The track gauge and pressure parameters collected during the placement process are monitored in each production cycle. The track gauge represents the center distance between the placement head and the track, while the pressure parameter refers to the contact force applied by the placement head when pressing down on the workpiece. Within a production cycle, the system records the track gauge and pressure values ​​corresponding to each placement action. For example, in cycle T1, the track gauge is recorded as 0.48mm and the pressure as 17.2N; in cycle T2, the track gauge is recorded as 0.50mm and the pressure as 18.1N. By comparing the parameter changes between the two cycles, it is calculated that an increase in track gauge from 0.48mm to 0.50mm and an increase in pressure from 17.2N to 18.1N both indicate a positive impact. If the trend is upward, meaning the direction is consistent, it indicates that there is a coordinated adjustment behavior between the mounting head position and the downward pressure in the current cycle. The parameter combination corresponding to this cycle is classified into the combination behavior in the upward direction. Further screening is performed, such as when the gauge decreases to 0.47mm and the pressure decreases to 16.8N in cycle T3, which is also regarded as a downward combination with the same direction. By continuously screening the combination behaviors with the same direction in multiple cycles, such as T1-T2, T3-T4, etc., they are recorded as parameter adjustment combination sequences. Each sequence can be represented as: cycle number + gauge direction + pressure direction, for example, T1: ↑↑, T3: ↓↓, finally obtaining the parameter adjustment combination sequence.

[0091] S202: Based on the parameter adjustment combination sequence, analyze the distribution of parameter change trends in adjacent production cycles, determine the relationship between each group of parameter adjustment behavior and the corresponding production cycle, mark the positions where parameter adjustments occur within the cycle as data structure inflection points, and obtain the structure inflection point sequence.

[0092] To determine the relationship between the adjustment behavior of each set of parameters and the corresponding production cycle, the following formula is used:

[0093] ;

[0094] Calculate the intensity value of parameter adjustment to determine and identify data structure inflection points;

[0095] in, Adjust the strength value of the parameters for the f-th device node. Let f be the normalized value of the pressure parameter of the f-th device node at period t. Let f be the normalized value of the pressure parameter of the f-th device node at period t+1. Let f be the normalized value of the track gauge parameter of the f-th device node at period t. This represents the normalized value of the track gauge parameter for the f-th device node at period t+1. Let f be the normalized value of the process stability coefficient of the f-th equipment node at period t. Let f be the normalized value of the process stability coefficient of the f-th equipment node at cycle t+1, where f is the equipment node index number, t is the current production cycle index number, and t+1 is the next adjacent production cycle index number.

[0096] In the above process of determining the relationship between the adjustment behavior of each set of parameters and the corresponding production cycle, the following formula is used to quantitatively describe the combined trend of changes in gauge and pressure parameters within adjacent cycles: The absolute difference in the numerator reflects the adjustment range of gauge and pressure parameters between two consecutive cycles. The denominator uses the normalized square root of the process stability coefficient within two cycles as a factor to suppress adjustment fluctuations. This formula simultaneously incorporates a trade-off between the joint changes in gauge and pressure and the process stability factor, ensuring the accuracy of identifying structural inflection points. The parameters are obtained as follows: Normalized pressure value. , The gauge normalized value is obtained by dividing the average pressure value collected by the pressure sensor in each cycle of the mounting process by the maximum pressure value of that process segment. , It is calculated by the ratio of the average track gauge recorded by the track gauge monitoring module to the track gauge design reference, and is the normalized value of the process stability coefficient. , This is the result after normalizing the scoring of the material feeding cycle stability in the mounting stability diagnostic module. The parameter assignment and calculation process is as follows: During the testing phase of a certain production line, by collecting the operating data of equipment A during cycles t and t+1, it was found that the average pressure within cycle t was 65 kPa, and the maximum pressure was 100 kPa. The average pressure within period t+1 is 80 kPa, and the maximum pressure is 100 kPa. The gauge value within period t is 1.75 mm, and the standard gauge is 5 mm. The gauge value within period t+1 is 1.50mm, and the standard gauge is 5mm. The beat stability score is 7 within period t, 9 within period t+1, and the maximum score is 10. Normalized, this yields... , ;

[0097] Substituting into the above formula, we get:

[0098] ;

[0099] Table 2. Normalization of Monitoring Parameters:

[0100]

[0101] As shown in Table 2, after normalization, all parameters are between 0 and 1, with consistent dimensions, which facilitates comprehensive calculations.

[0102] The results show that the parameter combination adjustment level between periods t and t+1 is 0.1581. This value can be compared with the structural turning point judgment reference value. When the structural turning point judgment benchmark is 0.12, this value is greater than the benchmark value, indicating that this is a data structure turning point, which should be recorded and subsequently packaged. By combining the gauge change, pressure change, and process stability factor into a unified calculation structure, the formula can realize the structural feature mining of parameter fluctuations at a single period scale, thus providing discriminative numerical support for anomaly identification and turning point marking in the SMT production process.

[0103] S203: Call the structural inflection point sequence, divide the production line data into multiple structural segments with the inflection point as the boundary, encapsulate the data according to the structural segment, establish structural segment data with structural segment number, encapsulation start time, and parameter direction combination code, and obtain the parameter structure encapsulation result;

[0104] The structural turning point sequence is invoked to segment the full-cycle data. For example, T1 to T3 is considered the first structural segment, T4 to T6 the second, T7 to T9 the third, and so on. The data within each structural segment is numbered according to its starting cycle time, such as segment 1 starting at 08:00:00, segment 2 starting at 08:07:45. The track gauge direction and pressure direction combination code within each segment is also defined, such as ↑↑ for the first segment, ↓↓ for the second, and ↑↓ for the third. This establishes the encapsulation identification information for each structural segment, which includes: structural segment number (e.g., SD01, SD02), starting time, direction combination code, cycle interval, and other data items. The encapsulated structural segment data is uniformly written into the structural segment management table. The table fields can be exemplified as: number = SD01, starting time = 08:00:00, cycle interval = T1-T3, direction combination = ↑↑. Multiple structural segment data records are generated in this way, ultimately obtaining the complete parameter structure encapsulation result.

[0105] Please see Figure 4 The steps to obtain the direction path combination sequence are as follows:

[0106] S301: Call the parameter structure encapsulation result, determine the direction of change of the mounting rate, visual offset and temperature control feedback in each structural segment, convert the change trend of each parameter into a direction code, and generate a parameter direction combination code by combining the three direction codes in each cycle.

[0107] The parameter structure encapsulation result is called, and the data changes of three parameters—placement rate, visual offset, and temperature control feedback—are sequentially obtained within each structure segment, using the structure segment as the boundary. For example, in structure segment SD01, the placement rate changes from 22.4 pieces / min to 24.1 pieces / min between cycles T1 and T3, the visual offset decreases from 0.03mm to 0.02mm, and the temperature control feedback increases from 198.3℃ to 200.5℃. The direction of change of the three parameters is determined to be ↑, ↓, and ↑, respectively, and the combination of these three directions is recorded as the direction code ↑↓↑. Similarly, in structure segment SD02, the placement rate decreases, the visual offset decreases, and the temperature control feedback decreases, which is uniformly coded as ↓↓↓. The parameters of all cycle groups within each structure segment are traversed, and the corresponding three-direction combinations are generated sequentially. The combination results are recorded in units of cycle number, such as ↑↓↑ for T3 and ↓↓↓ for T4. Finally, the parameter direction combination codes for all cycles within the structure segment are obtained.

[0108] S302: Based on parameter direction combination coding, identify combination path codes with consistent parameter change direction characteristics, count the frequency of occurrence of each group of direction combination codes in the corresponding structural segment, analyze the occurrence trend of each combination, and obtain the path code repetition frequency coefficient.

[0109] Based on the parameter direction combination codes generated for each segment, the frequency of direction codes appearing in all cycles is grouped and statistically analyzed. For example, ↑↓↑ appears 4 times in the SD01 structure segment, ↓↓↓ appears 2 times, and ↑↑↑ appears once. By accumulating the number of times each group of direction codes appears in the structure segment, the repetition frequency ratio of the combination in the current segment is calculated. The frequency ratio is expressed as the number of times the combination appears divided by the total number of cycles in the segment. If ↑↓↑ appears 4 times in a total of 7 cycles, its frequency coefficient is 0.571. The frequency coefficients of all combination codes are summarized to form a path statistics table and recorded by structure segment. The statistical results not only retain the combination code but also include the structure segment number, combination content, number of occurrences, and frequency coefficient. For example, the recording format is: SD01, combination ↑↓↑, number of occurrences 4, frequency coefficient 0.571. Further, based on the distribution of combination frequencies, it is analyzed which direction combinations constitute a repetition pattern in the segment, thereby obtaining the path code repetition frequency coefficient of the segment.

[0110] S303: Based on the path code repetition frequency coefficient, compare with the behavioral characteristic conditions of the dominant operating state path, filter the directional combination codes, mark the codes that meet the conditions as structural identifiers, organize them into path sequence numbers, establish the dominant behavioral combination of the periodic path, and obtain the directional path combination sequence.

[0111] Based on the path coding repetition frequency coefficient, combinations with a frequency coefficient greater than a certain standard value are screened out. For example, if the screening criterion is that the frequency coefficient in a single segment is not less than 0.45, then in segment SD01, ↑↓↑ is 0.571, which meets the condition, while ↓↓↓ is 0.286, which does not. Then, the behavioral characteristic conditions of the dominant operating state path are called, such as the combination of continuously increasing placement rate, stable temperature control, and decreasing offset as ↑↓–. Combinations that meet the frequency standard are then subjected to behavioral characteristic matching judgment. If the combination direction meets the behavioral condition, the combination code is recorded and marked as the structural identifier of the segment. Then, each matching combination code is assigned a unique path number, such as P001, P002, etc. The path numbers of matching combinations in different structural segments are arranged in the order of structural segments. Finally, the dominant path sequence corresponding to each structural segment of the entire production line is output, such as SD01–P001, SD02–P003, SD03–P001. After summarizing, the directional path combination sequence is obtained.

[0112] Please see Figure 5 The steps for obtaining the abnormal path signal identifier are as follows:

[0113] S401: Based on the direction path combination sequence, filter the recording time of abnormal signals in the mounting rate, feeding signal, and temperature control signal, analyze the triggering time sequence of the signals within the same cycle, cross-compare the start time and duration of each group of signals, and obtain abnormal signal time sequence distribution data.

[0114] Based on the dominant path number determined by the directional path combination sequence, such as path code P002 corresponding to structural segment SD04, the data records of placement rate, feeding signal, and temperature control signal in the corresponding periods T15 to T18 are retrieved. Recording periods with abnormal status changes are selected, such as a sudden drop in placement rate from 24.0 pieces / min to 18.7 pieces / min in T16, an intermittent feeding signal in T15, and a short-term temperature fluctuation exceeding 5°C in the temperature control signal in T17. A timeline coordinate system is established using the start point and duration of each signal recording time, forming an abnormal signal record table. Furthermore, the triggering sequence of each abnormal signal in T15–T18 is arranged, such as the material feeding abnormality triggering time being 08:33:17, the placement rate abnormality triggering time being 08:33:28, and the temperature control fluctuation triggering time being 08:34:02. These signal time sequences are organized and their duration is calculated, such as the material feeding abnormality lasting 47 seconds and the placement rate abnormality lasting 38 seconds. Based on the start time and duration range, the signal triggering sequence and time coverage are established, and finally, the abnormal signal time sequence distribution data in each cycle from T15 to T18 are obtained.

[0115] S402: Based on the time-series distribution data of abnormal signals, determine the triggering order between each signal, perform overlapping interval detection on the start time and duration of each group of signals, filter out signals that cover the response period of other signals in the time series, and obtain the signal type code of the coverage interval;

[0116] Based on the time-series distribution data of abnormal signals, the trigger order of the placement rate, feeding signal, and temperature control signal in each cycle is determined. For example, if a feeding anomaly triggers before a placement rate anomaly within T16, the determination method is to compare the start timestamps of each signal and then perform a cross-comparison of the time intervals formed by the start and end times of the signals. For instance, if the feeding signal trigger interval is from 08:33:17 to 08:34:04, and the placement rate anomaly interval is from 08:33:28 to 08:34:06, there is a significant overlap between the two. This overlap is then analyzed. As a basis for judging the coverage relationship, in the period where the three signals overlap, their respective durations and starting positions are compared to identify the signal type that completely covers the response periods of the other two signals in the time segment. Assuming that the temperature control signal has the longest duration and the earliest start in the T17 period, from 08:34:02 to 08:35:35, and all other signals are included in this segment, the temperature control signal is screened out as the main signal and assigned the signal type number W01. Finally, the corresponding coverage interval signal type code is obtained in all overlapping periods.

[0117] S403: Based on the signal type encoding of the coverage area, mark the main signal type selected in each group of periodic signals and use it as the starting signal of the main attribution chain. Organize the number, type, and periodic segment of the main signal, establish the type sequence of the main attribution chain, and obtain the abnormal path signal identifier.

[0118] Based on the signal type encoding of the coverage area, a marking operation is performed on the main signal types selected in each cycle. For example, the feeding signal in cycle T15 is the main signal, the mounting rate signal in cycle T16 is the main signal, and the temperature control signal in cycle T17 is the main signal. The signal number, corresponding type such as S01, S03, W01, and the cycle segment T15–T17 are recorded respectively, and a main signal number record table is established. At the same time, the starting cycle of each main signal type is used as the starting point of the chain, and the corresponding signals are sequentially connected to form an attribution chain path. The sequence of each attribution chain type is sorted out. For example, the attribution sequence under path P002 is S01→S03→W01, which successively represents the attribution development process of feeding abnormality, mounting rate abnormality, and temperature control abnormality. The result of this chain sequence is recorded and output to obtain the abnormal path signal identifier.

[0119] Please see Figure 6 The steps for obtaining the fault trend prediction results are as follows:

[0120] S501: Based on the abnormal path signal identifier, by analyzing the placement rate fluctuation amplitude, material supply stabilization time and temperature control response duration, calculate the equipment status score for each cycle and obtain the cycle status score sequence;

[0121] Based on the main attribution signals and their corresponding period segments marked in the abnormal path signal identifiers, such as the T21 to T24 segments which indicate placement rate fluctuations, material supply interruptions, and temperature control anomalies, the variation amplitude of the placement rate is statistically analyzed sequentially. For example, in period T21, the placement rate drops from 22.6 pieces / minute to 19.3 pieces / minute, with a fluctuation value of 3.3. The material supply signal interruption time is 45 seconds in period T22, and the temperature control response lag time reaches 58 seconds in period T23. The degree of change of the three parameters in each period is recorded, and the three values ​​are normalized. The scores are then calculated using fixed weighting coefficients. The weight for the placement rate fluctuation score is set to 0.4, the weight for the material supply stability score is set to 0.35, and the weight for the temperature control response score is set to 0.25. The score for each cycle is calculated as a weighted sum of the placement rate fluctuation score, the material supply time stability score, and the temperature control response time score. For example, in cycle T22, when the normalized values ​​of the three items are 0.65, 0.80, and 0.55, the corresponding cycle status score is 0.692. This calculation method is repeated to obtain the score values ​​for the four cycles from T21 to T24 and organize them to establish a cycle status score sequence.

[0122] S502: Based on the periodic state scoring sequence, analyze the step-by-step change process of continuous scoring results in each operating cycle, and statistically analyze the anomaly type attribution in each segment, identify the anomaly type path direction corresponding to each segment, and obtain the scoring change path distribution data.

[0123] Based on the above periodic status scoring sequence, the scoring values ​​of periods T21 to T24 are compared periodically to analyze whether the scoring results show an increasing, decreasing, or oscillating trend. For example, if the scoring sequence is 0.735, 0.692, 0.657, and 0.619, it shows a continuous downward trend. The downward path is recorded, and T21 to T24 are merged into a continuous scoring segment. Then, the abnormal signal types marked in this segment are statistically assigned. T21–T22 is mainly due to abnormal placement rate, T23 is mainly due to abnormal material supply, and T24 is mainly due to abnormal temperature control. This scoring segment corresponds to three abnormal types. The abnormal path direction is assigned according to their order in the scoring segment, and the path direction is constructed as rate → material supply → temperature control. After repeating this process for different continuous scoring segments, the trend direction and abnormal attribution corresponding paths of all scoring segments are sorted out to obtain the scoring change path distribution data.

[0124] S503: Based on the distribution data of the scoring change path, determine the migration path direction corresponding to the anomaly type in multiple scoring change segments, identify the migration path code and fault type number under each operating cycle, predict the fault evolution path of the equipment, and generate fault trend prediction results.

[0125] Based on the distribution data of scoring change paths, the correspondence between the scoring trend structure and abnormal path direction of multiple scoring segments is identified. For example, a continuously decreasing scoring segment often corresponds to an abnormal combination path at the beginning of the mounting rate. The path trend direction is combined with the attribution type number to generate a migration path code. For example, the path rate → feeding → temperature control corresponds to code P105. At the same time, its frequency of occurrence and segment span in the cycle sequence are recorded. When the same path code exists in multiple segments, its cumulative number is recorded as an important indicator of the equipment operation evolution trend. Then, the code is compared with the pre-set fault type dictionary. For example, if the path direction of P105 corresponds to the feeding component failure type F07 in the mounting process, the path is matched with the fault type F07, and its occurrence cycle and inferred trend direction are recorded. Finally, the potential evolution paths of each type of fault in each cycle are integrated, and the fault trend prediction results of the equipment under the monitoring cycle are output.

[0126] Please see Figure 7 The SMT production line analysis system based on real-time data acquisition includes:

[0127] The node connection identification module calls the multi-device operation log to determine the time sequence of the completion and start signals of multiple devices, calculates the time interval between the signals and compares it with the communication beat standard, and identifies the connection relationship between devices based on the time sequence and time interval, generating device connection structure information.

[0128] The structural data segmentation module uses equipment connection structure information to analyze the direction of change of gauge and pressure parameters in adjacent cycles during the mounting process, identify parameter adjustment behavior, and mark it as a data structure turning point, dividing the data into multiple structural segments and generating parameter structure encapsulation results.

[0129] The path behavior coding module calls the parameter structure encapsulation results, compares the changing directions of mounting rate, visual offset, and temperature control feedback, identifies the combined path codes with the same direction, analyzes the repetition frequency, judges the dominant running path behavior characteristics, and generates a directional path combination sequence.

[0130] The dominant path identification module uses the directional path combination sequence to analyze the recording time of multiple abnormal signals, compares the signal triggering order and duration, determines the response cycle coverage relationship, records the directional combination according to the dominant path behavior characteristic conditions, and generates abnormal path signal identifiers.

[0131] The behavior trend prediction module analyzes the placement rate fluctuation, material supply stabilization time and temperature control response time based on the abnormal path signal identifier, calculates the cycle status score, analyzes the relationship between score changes and abnormal type attribution, matches the migration path direction, predicts the equipment failure evolution path, and generates failure trend prediction results.

[0132] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for SMT production line analysis based on real-time data acquisition, characterized in that, The method comprises the following steps: S1: calling a multi-device running log, judging the time sequence of a plurality of device running completion signals and start signals, calculating the time interval between the signals and comparing the communication beat standard, identifying the connection relationship between the devices according to the time sequence and the time interval, and generating device connection structure information; S2: using the mounting process node identified by the device connection structure information, calling and analyzing the change direction of the track pitch and pressure parameters of the corresponding node within the adjacent period, identifying the parameter adjustment behavior, marking the data structure turning point, dividing the data into a plurality of structure segments, and generating a parameter structure packaging result; S3: calling the parameter structure packaging result, comparing the change direction of the mounting rate, visual offset and temperature control feedback, identifying the combined path code with the same direction, analyzing the repetition frequency, judging the dominant running path behavior characteristics, and generating a direction path combination sequence; S4: using the direction path combination sequence, analyzing the recording time of a plurality of abnormal signals, comparing the signal trigger sequence and the duration section, judging the response period coverage relationship, and recording the direction combination according to the dominant path behavior characteristic condition to generate an abnormal path signal identifier; The acquisition step of the abnormal path signal identifier is: S401: according to the direction path combination sequence, screening the recording time of abnormal signals in the mounting rate, feeding signal and temperature control signal, analyzing the trigger time sequence of the signals in the same period, and cross-comparing the start time and duration of each group of signals to obtain abnormal signal time sequence distribution data; S402: based on the abnormal signal time sequence distribution data, judging the trigger sequence between each signal, performing overlap interval detection on the start time and duration of each group of signals, screening the signals covering the response periods of the remaining signals in the time sequence to obtain an overlap interval signal type code; S403: according to the overlap interval signal type code, marking the main signal type selected in each group of period signals as the starting signal of the main attribution chain, arranging the number, type and period section of the main signal, establishing the type sequence of the main attribution chain, and obtaining the abnormal path signal identifier.

2. The SMT line analysis method based on real-time data collection according to claim 1, characterized in that, The device connection structure information includes node connection sequence, time interval between nodes and connection path direction, the parameter structure packaging result includes structure segment number, packaging start time and parameter change direction combination, the direction path combination sequence specifically refers to direction code type, path combination number and repetition frequency in the period, and the abnormal path signal identifier specifically refers to abnormal signal start sequence, coverage period range and main attribution signal type.

3. The SMT line analysis method based on real-time data collection of claim 1, wherein, The communication beat standard refers to the signal response interval specification that should be met between a plurality of devices in the cooperative running process, reflecting the standard time configuration between the device completion signal and the start signal of the lower-level device; The dominant path behavior characteristic condition refers to the high-frequency combination trend of the three types of state characteristics of mounting rate, visual offset and temperature control feedback in the running period, the repetition frequency of the direction code is counted and its duration period is identified, and the direction path combination with the repetition frequency higher than the preset frequency threshold and the duration period longer than the preset period threshold is defined as the dominant path.

4. The SMT line analysis method based on real-time data collection of claim 1, wherein, The device connection structure information acquisition step is: S101: Call the multi-device running log, collect the time information of the start signal and the completion signal of each device in the SMT production line, including the placement machine, reflow soldering, and printing machine, sequentially sort and pair the time information, analyze the time sequence of device operation, and generate a device signal sequence; S102: Based on the device signal sequence, calculate the time interval between the start signal and the completion signal of adjacent devices, compare the time interval with the communication beat standard, establish the connection path between nodes, and obtain the connection relationship coefficient between nodes; S103: According to the connection relationship coefficient between nodes, the connection sequence between each device node, the corresponding time interval of the node, and the connection direction are summarized, the running process topology is constructed, and the device connection structure information is obtained.

5. The SMT line analysis method based on real-time data collection according to claim 4, characterized in that, The parameter structure packaging result acquisition step is: S201: Using the mounting process node identified by the device connection structure information, call and analyze the track spacing and pressure parameters of the corresponding node, and calculate the change state in each production cycle. The track spacing parameter change value and the pressure parameter change value of each cycle are executed. Direction judgment, filter parameter adjustment behaviors with consistent change direction in the same cycle, and obtain parameter adjustment combination sequence; S202: Based on the parameter adjustment combination sequence, analyze the distribution of parameter change trend in adjacent production cycles, judge the relationship between each group of parameter adjustment behaviors and the corresponding production cycle, mark the position of parameter adjustment in the cycle as a data structure turning point, and obtain a structure turning point sequence; S203: Call the structure turning point sequence, divide the production line data into multiple structure sections with the turning point as the boundary, perform data packaging according to the structure section, establish the structure section number, packaging start time, and parameter direction combination code structure section data, and obtain the parameter structure packaging result.

6. The SMT line analysis method based on real-time data collection according to claim 5, characterized in that, The direction path combination sequence acquisition step is: S301: Call the parameter structure packaging result, perform change direction judgment on the mounting rate, visual offset, and temperature control feedback in each structure section, convert the change trend of each parameter into a direction code, and combine the three direction codes of each cycle to generate a parameter direction combination code; S302: Based on the parameter direction combination code, identify the combination path code with consistent parameter change direction characteristics, count the occurrence frequency of each group of direction combination codes in the corresponding structure section, analyze the occurrence trend of each combination, and obtain the path code repetition frequency coefficient; S303: According to the path code repetition frequency coefficient, compare the behavior characteristic conditions of the dominant running state path, filter the direction combination codes, mark the codes that meet the conditions as structure identifiers, arrange them into path sequence numbers, establish the dominant behavior combination of the cycle path, and obtain the direction path combination sequence.

7. The real-time data acquisition based SMT line analysis method according to claim 1, characterized in that, The method further comprises: S5: According to the abnormal path signal identifier, analyze the mounting rate fluctuation, feeding stabilization time, and temperature control response time, calculate the cycle state score, analyze the score change and abnormal type attribution relationship, match the migration path direction, predict the fault evolution path of the device, and generate a fault trend prediction result; The fault trend prediction result is specifically a score change section identifier, an abnormal migration path number, and a fault type prediction result.

8. The SMT line analysis method based on real-time data collection according to claim 7, characterized in that, The obtaining of the fault trend prediction result comprises the following steps: S501: According to the abnormal path signal identifier, the device state score of each period is calculated by analyzing the mounting rate fluctuation amplitude, the feeding stabilization time and the temperature control response duration, and the period state score sequence is obtained; S502: Based on the period state score sequence, the step-by-step change process of the continuous score result in each running period is analyzed, the abnormal type attribution in each section is counted, the abnormal type path direction corresponding to each section is identified, and the score change path distribution data is obtained; S503: According to the score change path distribution data, the migration path direction corresponding to the abnormal type in multiple score change sections is judged, the migration path code and the fault type number under each running period are identified, the fault evolution path of the device is predicted, and the fault trend prediction result is generated.

9. A SMT production line analysis system based on real-time data acquisition, characterized by, The system is used to realize the SMT production line analysis method based on real-time data collection according to any one of claims 1-8, and the system comprises: The node connection identification module calls the multi-device running log, judges the time sequence of the multi-device running completion signal and the start signal, calculates the time interval between the signals and compares the communication beat standard, identifies the connection relationship between the devices according to the time sequence and the time interval, and generates the device connection structure information; The structure data segmentation module uses the device connection structure information to analyze the change direction of the track pitch and the pressure parameter in the adjacent period in the mounting process, identifies the parameter adjustment behavior, and marks it as a data structure turning point, divides the data into multiple structure sections, and generates the parameter structure packaging result; The path behavior coding module calls the parameter structure packaging result, compares the change direction of the mounting rate, the visual offset and the temperature control feedback, identifies the combined path code with the same direction, analyzes the repetition frequency, judges the dominant running path behavior characteristics, and generates the direction path combination sequence; The dominant path identification module uses the direction path combination sequence to analyze the recording time of the multiple abnormal signals, compares the signal trigger sequence and the continuous section, judges the response period coverage relationship, records the direction combination by comparing the dominant path behavior characteristic condition, and generates the abnormal path signal identifier; The behavior trend prediction module analyzes the mounting rate fluctuation, the feeding stabilization time and the temperature control response time according to the abnormal path signal identifier, calculates the period state score, analyzes the score change and the abnormal type attribution relationship, matches the migration path direction, predicts the fault evolution path of the device, and generates the fault trend prediction result.

Images