SMT production line equipment fault diagnosis method and system based on Internet of Things

Abstract

The invention relates to the technical field of fault diagnosis, in particular to an SMT production line equipment fault diagnosis method and system based on the Internet of Things, and the method comprises the following steps: obtaining multiple types of signals, aligning a time window, judging that the trend is consistent and fluctuation is synchronous to generate a sudden change event, recognizing an intersection point, extracting a key response position, and outputting an inflection point node number. And constructing a response time sequence and a propagation path, identifying abnormal nodes, and generating a fault traceability result. According to the invention, through time alignment and trend linkage identification of current, temperature and acceleration signals, fusion of trend direction judgment and fluctuation synchronization relation, a cross-equipment linkage sudden change identification mode is established, key response points are extracted through cross positioning of a jump terminal point and a vibration peak value, and a dynamic judgment mechanism of inflection point nodes is formed. In combination with a response sequential sequence and a chain path structure, local signal fluctuation analysis is converted into perception, and the judgment accuracy and traceability of abrupt change abnormity are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault diagnosis, in particular to a SMT production line equipment fault diagnosis method and system based on Internet of Things. BACKGROUND

[0002] The technical field of fault diagnosis involves monitoring and analyzing the running state of industrial systems, mechanical equipment and automated production lines, etc. Its core tasks include data collection of equipment running state, abnormal state identification, fault mode classification, fault location judgment and fault cause inference, etc. This technical field systematically covers monitoring methods and analysis methods for the life cycle of equipment, and plays an important role in ensuring stable operation of equipment, improving production efficiency and reducing unplanned downtime, especially in manufacturing. In modern manufacturing scenarios, fault diagnosis technology is widely used in numerical control machine tools, automated assembly lines, electronic manufacturing lines and other fields, and gradually evolves from manual experience judgment to intelligent diagnosis based on sensor data. Among them, the traditional SMT production line equipment fault diagnosis method refers to the state judgment and problem troubleshooting of pick-and-place machines, reflow soldering, printing machines and other key equipment through the experience judgment of operators or the offline analysis of single sensor data. This method mainly relies on temperature detection devices, current monitoring circuits, vibration sensors and other single-point physical quantity acquisition devices, and then analyzes the abnormality through manual analysis or preset threshold judgment. In the fault identification of reflow soldering equipment, the traditional method usually judges the abnormality according to the deviation degree of the heating time and temperature curve of the hot zone; in the diagnosis of pick-and-place machines, the fluctuation range of motor driving current is used as the basis for the state change of transmission components; in the state identification of printing machines, the mode matching method based on vibration signal amplitude and frequency change is widely used to identify the problems of scraper loosening or deviation. This traditional method relies on fixed detection points and single signal channels, and it is difficult to form a overall perception of the running state of the whole production line.

[0003] The existing technology relies on the experience judgment of operators or the static analysis of single sensor data, and it is difficult to capture the linkage abnormal signals between devices in the multi-device collaborative running scenario. Especially in the SMT production line, the running states of pick-and-place machines, reflow soldering and printing machines interact frequently. If only single type of signal is collected through fixed detection points for analysis, the interaction trend between signals is easily ignored, leading to lag in identifying sudden abnormal events. For example, when the temperature of the hot zone of the reflow soldering fluctuates rapidly, if the current of the pick-and-place machine also rises suddenly, single device analysis cannot capture the cooperative abnormal change of the cross-device signal, resulting in delay in fault location and failure in abnormal source tracing. SUMMARY

[0004] In order to solve the technical problems existing in the prior art, the embodiments of the present application provide a SMT production line equipment fault diagnosis method based on Internet of Things, which comprises the following steps: In order to achieve the above object, the present application adopts the following technical scheme: A SMT production line equipment fault diagnosis method based on Internet of Things, comprising the following steps: S1: Obtain the current signal of the SMT production line chip mounter driving motor, the temperature signal of the reflow solder heating area and the acceleration signal of the printing machine doctor blade endpoint, perform trend direction judgment on the current jump-up section, the temperature continuous rise section and the acceleration fluctuation section, and generate equipment node linkage mutation event identification information; S2: Based on the equipment node linkage mutation event identification information, call the current signal jump-up section position and the acceleration signal waveform peak position, cross-position the jump-up end and the vibration peak, identify whether the waveforms of the two are time-coincident in a continuous period, obtain the inflection point equipment node number record; S3: According to the inflection point equipment node number record, collect the first response time stamp of the chip mounter, reflow solder and printing machine equipment in the mutation period, number sort according to the SMT production line sequence, eliminate the nodes without response data, and obtain the initial state response time sequence; S4: Call the initial state response time sequence, perform connection judgment on the time sequence of the SMT production line equipment, construct a series structure starting from the first node, form a chain arrangement response path, and obtain chain control response structure path data.

[0005] As a further scheme of the present application, the equipment node linkage mutation event identification information includes trend consistency judgment result, signal fluctuation synchronization and time window alignment information, the inflection point equipment node number record includes mutation trend key response position number, equipment identification information and trend variation point label, the initial state response time sequence includes response equipment number, first response time stamp and number sorting result, and the chain control response structure path data includes first node, propagation sequence and response path chain arrangement structure.

[0006] As a further scheme of the present application, the specific steps of S1 are: S101: Obtain the current signal of the SMT production line chip mounter driving motor, the temperature signal of the reflow solder heating area and the acceleration signal of the printing machine doctor blade endpoint, combine the uniform sampling period to synchronize and align the time sequence of the three types of signals, call the original time stamp of the signal and match the corresponding sampling point number, construct a time window according to the same number time point, and obtain a synchronous period signal sequence group; S102: According to the chip mounter driving current sequence, the reflow solder heating area temperature sequence and the doctor blade endpoint acceleration sequence in the synchronous period signal sequence group, extract the value interval of the current jump-up section, the temperature continuous rise section and the acceleration fluctuation section in the time window, identify the signal amplitude change trend from the start to the end of the interval section, and obtain the trend consistency direction combination; S103: Based on the trend consistency direction combination, the fluctuation frequency of the signals in the combination is compared in sequence, the number of fluctuation peaks and the interval time in the corresponding time window are extracted, and it is judged whether the fluctuation section fluctuation timing coincidence condition is met, if the determination result is true, the fluctuation coincidence interval of the signal combination is extracted and the time node position is marked, and the device node linkage mutation event identification information is obtained.

[0007] As a further scheme of the application, the specific steps of S2 are: S201: Based on the device node linkage mutation event identification information, the position of the current signal in the corresponding jump-up section and the position of the waveform peak in the acceleration signal are called, the jump-up section end time point and the peak time point are extracted, comparison is made according to the same time axis number, it is screened whether the two exist intersection in the same time section range in the continuous period, if there is intersection, the time point is marked, and the jump-up peak intersection time point set is obtained; S202: According to the time of the intersection point in the jump-up peak intersection time point set, the position of the intersection point in the original signal sequence is extracted, the trend section starting position of the current signal and the acceleration signal is searched in sequence, the leading node position of the intersection point in the continuous trend section is marked, the node number is called for labeling, and the trend section leading node index group is obtained; S203: The node number marked in the trend section leading node index group is called, the fluctuation density of the current signal and the acceleration signal in the front and rear time sections is extracted, the fluctuation frequency correction value is calculated, it is judged whether the trend dense section synchronization feature is presented in the section, if the frequency range is consistent and the interval difference is within the limited interval, the node is marked as a key response point, and the mutation trend key response node number set is obtained; S204: Based on the device identification code corresponding to the nodes in the mutation trend key response node number set, the device number information is extracted and renumbered, the nodes are sorted according to the original index sorting mode in the signal sequence, the marked node device number is output, and the inflection point device node number record is obtained.

[0008] As a further scheme of the application, the fluctuation frequency correction value is a quantitative value determined by calculating the first difference or variance of the fluctuation time interval.

[0009] As a further scheme of the application, the specific steps of S3 are: S301: According to the device number list in the inflection point device node number record, the chip mounter, reflow soldering and printer device types are classified and processed respectively, the signal sequence of each type of device in the mutation period is called, the first fluctuation mutation position in the corresponding signal of the device is searched, the time stamp corresponding to the position is recorded, and the device first response time stamp group is obtained. S302: Based on the timestamps and equipment numbers in the equipment first response timestamp group, the equipment numbers are prioritized according to the process sequence of printer, pick-and-place machine and reflow soldering in the SMT production line. The equipment response deviation value is obtained by calculating the difference between the first response timestamp of each equipment and the expected response timestamp based on the process sequence. Equipment numbers with deviation values ​​exceeding a preset threshold are removed. The remaining equipment numbers are paired with their corresponding first response timestamps to obtain a sequence matching response record group. S303: Call the combination value of the device number and timestamp in the sequence matching response record group, and splice them according to the sorted process order and time sequence to obtain the initial state response timing sequence.

[0010] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Call the device number and corresponding timestamp data in the initial state response timing sequence, compare them item by item from early to late according to the timestamp, and determine whether the time sequence between adjacent devices is continuously increasing. If the increasing condition is met, establish a connection relationship between adjacent device numbers, and add them sequentially according to the time order to obtain the first and last connected node path. S402: Based on the sequence of device numbers that have been connected in the first and last connected node paths, the starting node and the ending node are marked, and the propagation position number of each node in the structure is marked sequentially from the first node to obtain the chain control response structure path data.

[0011] As a further aspect of the present invention, the method further includes the following steps: S5: Using the path data of the chain control response structure, determine whether the continuous rise in temperature signal and the sudden rise in current of the first node device have an overlapping trend, analyze whether it has the characteristics of multiple signal types of sudden combination, and identify whether the node is an abnormal initiating node in the chain control structure based on whether the signal starting waveform and the node waveform have a synchronous change trend in the previous cycle, and generate the fault tracing device node identification result. The fault tracing device node identification results include the anomaly initiation node identifier, mutation combination feature matching results, and starting waveform synchronization determination label.

[0012] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Call the first node device number in the chain control response structure path data, obtain the temperature signal sequence and current signal sequence corresponding to the device, extract the continuous rising segment of the temperature signal and the sudden rising segment of the current signal within the mutation cycle, compare them according to the same time axis, and determine whether the two signals appear synchronously in adjacent or overlapping segments. If they appear synchronously, they are marked as overlapping trends, and the mutation mode overlap flag value is obtained. S502: Based on the initial node device number corresponding to the true value of the mutation mode overlap flag, extract the waveform of the initial segment of temperature and current signals in the previous cycle, compare it with the waveform corresponding to the node in the mutation interval, and make a synchronization judgment according to the waveform slope direction and change density to obtain the synchronization trend conforming node index set. S503: Call the node number marked in the synchronization trend matching node index set, and combine it with the sorting position in the chain control response structure path data to filter the first node that simultaneously meets the characteristics of overlapping mutation mode and trend synchronization. Extract the device identification code of the first node, mark it separately, and obtain the fault tracing device node identification result.

[0013] An IoT-based SMT production line equipment fault diagnosis system includes: The signal acquisition module acquires the current signal of the pick-and-place machine, the temperature signal of the reflow soldering machine, and the acceleration signal of the printer. After unifying the sampling period, the timestamps of the signals are normalized to generate identification information of sudden events in the linkage of equipment nodes. Based on the device node linkage sudden event identification information, the trend linkage identification module calls the current signal jump section position and the acceleration signal waveform peak position to cross-locate the jump endpoint and vibration peak, identify whether the two waveforms overlap in time within a continuous period, and generate inflection point device node number record; The initial response sequence generation module collects the first response timestamps of the pick-and-place machine, reflow soldering machine and printing machine during the mutation cycle according to the inflection point equipment node number record, sorts the numbers according to the SMT production line sequence, removes nodes without response data, and generates the initial state response time sequence. The response path construction module determines the connection relationship between the time sequence of SMT production line equipment based on the initial state response timing sequence, constructs a series structure starting from the first node, forms a chain-arranged response path, and obtains the chain-controlled response structure path data. The abnormal node identification module determines the connection relationship between SMT production line equipment based on the initial state response timing sequence, constructs a serial structure starting from the first node, forms a chain-like response path, and obtains the chain control response structure path data.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by time alignment and trend linkage identification of multiple signals such as current, temperature and acceleration, and by integrating trend direction judgment and fluctuation synchronization relationship, a linkage mutation identification method between devices is established. By cross-location of jump endpoint and vibration peak value to extract key response points, a dynamic judgment mechanism for inflection point nodes is formed. By combining response time sequence and chain path structure, a device response sequence chain is constructed to capture the synchronous mutation characteristics of multi-signal fusion features in the propagation path, thereby realizing the effective identification of abnormal initiation nodes and fault tracing, completing the transformation from local signal fluctuation analysis to perception, and effectively improving the accuracy and traceability of mutation anomaly judgment. Attached Figure Description

[0015] 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.

[0016] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention; Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] 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.

[0019] 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.

[0020] 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.

[0021] 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.

[0022] Please see Figure 1 This invention provides a method for fault diagnosis of SMT production line equipment based on the Internet of Things, including the following steps: S1: Obtain the current signal of the drive motor of the SMT production line pick and place machine, the temperature signal of the reflow soldering heating zone and the acceleration signal of the printing machine squeegee tip. Based on the same sampling period, align the time windows of the three types of signals and perform trend direction judgment on the current jump segment, the temperature continuous rise segment and the acceleration fluctuation segment. If any two trends are consistent and the fluctuation occurs at the same time, generate the equipment node linkage sudden change event identification information. S2: Based on the device node linkage sudden change event identification information, call the current signal jump section position and the acceleration signal waveform peak position, cross-locate the jump end point and vibration peak, identify whether the two waveforms show time overlap within a continuous period. If overlap occurs, mark it as the intersection point. Identify the position at the leading edge of the continuous trend segment from the intersection point, and determine whether the dense fluctuation trend of the preceding and following signals is synchronized. If synchronized, mark the intersection point as the key response position of the sudden change trend, and output the corresponding device number as the inflection point node to obtain the inflection point device node number record; S3: Based on the inflection point equipment node number record, collect the first response timestamp of the pick-and-place machine, reflow soldering machine and printing machine during the mutation cycle, sort the numbers according to the SMT production line sequence, remove nodes with no response data, and combine the remaining numbers with the time to obtain the initial state response timing sequence. S4: Call the initial state response timing sequence to determine the connection relationship between the time sequence of SMT production line equipment, construct a series structure starting from the first node, form a chain-like response path, mark the propagation order between the first node and the last node, and obtain the chain control response structure path data. S5: Using the chain control response structure path data, determine whether the continuous rise in temperature signal and the sudden rise in current of the first node device have an overlapping trend, analyze whether it has the characteristics of multiple signal types of sudden combination, and identify whether the node is an abnormal initiating node in the chain control structure based on whether the signal starting waveform and the node waveform have a synchronous change trend in the previous cycle, and generate the fault tracing device node identification result. The device node linkage mutation event identification information includes trend consistency judgment results, signal fluctuation synchronization, and time window alignment information. The inflection point device node number record includes the key response position number of the mutation trend, device identification information, and trend variation point label. The initial state response time sequence includes the responding device number, the first response timestamp, and the number sorting result. The chain control response structure path data includes the first node, the propagation order, and the chain arrangement structure of the response path. The fault source tracing device node identification results include the abnormal initiation node identifier, mutation combination feature matching results, and the starting waveform synchronization judgment label.

[0023] Please see Figure 2 The specific steps of S1 are as follows: S101: Acquire the current signal of the drive motor of the SMT production line pick and place machine, the temperature signal of the reflow soldering heating zone, and the acceleration signal of the squeegee tip of the printer. Combine the time series of the three types of signals with a unified sampling period to perform synchronization and alignment processing. Call the original timestamp of the signal and match the corresponding sampling point number. Construct a time window based on the time points with the same number to obtain the signal sequence group of the synchronization period. Under a unified sampling period, the original sensor data of the three types of equipment must first be read. The current signal of the pick-and-place machine can be read through the current detection channel of the drive controller to read the original data log. For example, the encoder feedback interface built into the Omron driver can be used to read the A and B phase current values. The temperature signal of the reflow soldering heating zone is connected to the temperature control module through a K-type thermocouple to read the real-time temperature. The acceleration signal of the squeegee tip of the printer is acquired by a MEMS triaxial accelerometer (such as ADXL345) fixed on the squeegee assembly to obtain the data sequence. The acceleration values ​​are stored in the local acquisition buffer in the X, Y, and Z directions. The original data of the three types of signals are all timestamped. The equipment needs to be set to a unified sampling period, such as 10ms. During the reading process, a synchronous clock mechanism is used to align the data of each type according to the timestamp. If the sampling frequency of a certain type of signal is higher than 10ms, the moving average method or downsampling is used to process it. If there are missing sampling points in a certain type of signal, the interpolation method is used to fill in the missing points to obtain a synchronous time period signal sequence group.

[0024] S102: Based on the pick-and-place machine drive current sequence, reflow soldering heating zone temperature sequence, and scraper end acceleration sequence in the synchronous time period signal sequence group, extract the value ranges of the current jump segment, the temperature continuous rise segment, and the acceleration fluctuation segment within the time window, identify the signal amplitude change trend from the beginning to the end of the interval segment, and obtain the trend consistency direction combination. Extracting the current jump segment of the pick-and-place machine current signal requires traversing the sampling point values ​​within each time window. The presence of a jump is determined by comparing the differences between adjacent sampling points. The criterion is that the difference between three consecutive sampling points is greater than a set jump threshold, such as 0.5A, to mark it as a jump segment. If a data segment is 2.0A, 2.6A, 3.1A, and 3.5A, then that interval can be extracted as a current jump segment. During the extraction of the temperature rise segment of the reflow soldering heating zone temperature signal, the total change in temperature within any continuous interval is calculated. When the change is greater than 2.0℃ and the duration exceeds 500ms (corresponding to 50 sampling points), it can be considered a continuous temperature rise segment. Extraction of the acceleration signal fluctuation segment requires first calculating the total change in temperature within each time window. The standard deviation of acceleration data within the window is considered to have significant fluctuations when it exceeds a set fluctuation threshold, such as 0.3 m / s². After the above extraction is completed, the signal change trend in each segment is determined, i.e., the overall increase or decrease direction from the initial value to the final value. If the current increases from 2.1A to 3.6A, the temperature increases from 195℃ to 202℃, and the acceleration fluctuation value increases from ±0.2 m / s² to ±0.5 m / s², it is considered to be enhanced. If the three types of signals have the same direction, they are classified into the same trend combination. The trend direction is marked in all time windows to form a trend consistency direction combination. All types of signals in window number W8 are set to be in the upward direction, which is the trend consistency combination of "increase-increase-enhancement".

[0025] S103: Based on the trend consistency direction combination, compare the frequency of signal changes in the combination in turn, extract the number of fluctuation peaks and valleys and the interval time interval within the corresponding time window, and determine whether the fluctuation segment fluctuation time sequence overlap condition is met. If the judgment result is true, extract the fluctuation overlap interval of the signal combination and mark the time node position to obtain the device node linkage sudden change event identification information. In each window, the fluctuation characteristics of the three types of signals are analyzed sequentially. A three-point sliding window is used to detect extreme points for each signal sequence. If a point is greater than the points before and after it, it is a peak; if it is less, it is a trough. By iterating through the sequence, the number of peaks and troughs for each signal within the current time window and their corresponding sampling point positions are obtained. The fluctuation frequency is obtained by counting the number of extreme points. The calculation method is that the fluctuation frequency equals the number of extreme points divided by the window duration. If there are 4 current peaks and the window length is 1 second, the frequency is 4Hz. Then, the frequency difference of the three types of signals is compared. If the frequency error between the three is within 10%, then... If the current is 4Hz, the temperature is 4.2Hz, and the acceleration is 3.8Hz, then the condition of consistent fluctuation frequency is met. The time interval between extreme points is calculated, the corresponding extreme value index sequence is extracted, and the interval value list between adjacent extreme points is calculated using difference. The standard deviation is calculated. If the standard deviation is less than a certain set value, such as 20ms, then the fluctuation time is determined to have synchronous characteristics. The time window number of the signal combination that meets the two conditions of similar fluctuation frequency and coincidence of extreme point time is extracted, and the starting time point of the window is used as the identifier of the linkage mutation event. The linkage mutation event identifier information of the device node is obtained.

[0026] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the device node linkage sudden event identification information, call the position of the current signal in the corresponding jump section and the position of the waveform peak in the acceleration signal, extract the end time point and peak time point of the jump section, compare them according to the same time axis number, and filter whether the two have an intersection in the same time period within the continuous period. If there is an intersection, mark the time point to obtain the jump peak intersection time point set. Based on the device's previously identified records of device-linked sudden events, the device will call the current signal jump segment and the acceleration signal peak position within the time window corresponding to the event. By using a unified sampling time number, the last point of the jump end in the current signal is extracted as the jump endpoint. Then, the local maxima points within the same segment of the acceleration signal are extracted as waveform peaks. The time numbers of the two signals are placed into the time axis structure for item-by-item matching. The overlap of time numbers between the jump endpoint and the peak is compared. If there is a common number within an adjacent time number range (e.g., ±2 points), it is considered that there is an intersection relationship. Cross-matching operations are performed in multiple consecutive sampling windows to extract the intersection situation. The device aggregates the time number pairs into a unified data structure, representing the situation where the current jump and the acceleration peak are close to or superimposed on each other within the same time segment, thus obtaining the set of jump peak intersection time points.

[0027] S202: Based on the time of the intersection of the peak jump time points, extract the position of the intersection point in the original signal sequence, sequentially search forward to the starting position of the trend segment of the current signal and acceleration signal, mark the position of the leading node of the intersection point in the continuous trend segment, call the node number for labeling, and obtain the trend segment leading node index group. The actual sampling position corresponding to the point in the current and acceleration sequences is obtained from the raw data. Using this position as a reference starting point, the continuous trend of the sampling points is retrieved step by step forward in the sequence. In the current signal, the current point is used as the reference to compare the difference direction between adjacent values ​​point by point to identify the starting point of the upward change. If several consecutive upward slopes are set, the point where the first upward value is located is taken as the starting point of the trend. The same method is used to analyze the fluctuation trend between adjacent values ​​in the acceleration signal. The local change direction is compared to see if there is a continuous amplification or contraction process to identify the initial stage of the trend. If three consecutive points constitute a significant change segment, it is considered to be the beginning of the leading edge of the fluctuation trend. The trend starting points obtained by backtracking the current and acceleration signals are recorded in the time series. The number constitutes the leading edge node of the trend segment. By backtracking the convergence point of the intersection time point and performing the above operation in sequence, the device collects the leading edge starting points into the trend segment leading edge node index group in chronological order.

[0028] S203: Call the node number marked in the trend segment leading edge node index group, extract the fluctuation density of current signal and acceleration signal in the preceding and following time periods, calculate the fluctuation frequency correction value, and determine whether the trend dense segment synchronous characteristics are present in the segment. If the frequency range is consistent and the interval difference is within the limited range, then mark the node as a key response point and obtain the set of key response node numbers for sudden trend. The fluctuation frequency correction value is calculated using the following formula: ; in, Representing the The frequency correction value for the type of signal. Representing the Class of signals The length of the time interval within each fluctuation range Representing the The average length of the peak-valley interval of the signal. Representing the The sum of squares of the time intervals between the fluctuation ranges of the signal. This represents the total number of peak-valley intervals. , It is a serial number index variable; The calculation logic of the formula is as follows: By calculating the absolute value of the deviation between each time interval and the overall average time interval, and combining it with the normalized proportion of that time interval in the overall time distribution, a product is formed to reflect the fluctuation intensity of that interval in the whole signal. The products of the intervals are summed as the weighted accumulation of the total fluctuation amplitude. By normalizing the overall segment time, the frequency of fluctuation change per unit time is obtained. The formula integrates multiple calculation methods such as absolute difference, normalization and weighted summation, taking into account the overall characteristics of signal fluctuation density and time sequence distribution, so that the output not only reflects the frequency of change, but also describes the relative stability of fluctuation intensity. The fluctuation frequency correction value measures the degree of change in the fluctuation time interval of a signal over a period of time. It takes into account the difference between each fluctuation and the overall average interval and its relative proportion. The larger the value, the more frequent and uneven the fluctuations are, while the smaller the value, the more uniform and stable the fluctuations are. Parameter acquisition and quantization process description: A set of current signal (unit: A) fluctuation interval data within 10 seconds was selected, with a sampling frequency of 1000Hz. The fluctuation interval was extracted by setting the peak-valley threshold difference method, and the following 5 time intervals (seconds) were collected: ; Calculate the average time interval: ; Calculate the sum of squares: ; Calculate the square root: ; The calculations are as follows: Table 1: Weighted Processing Table for Experimental Data

[0029] As shown in Table 1, substitute the above data into the formula;

[0030] Calculate the sum of the numerators: ; Calculate the denominator: ; Substitute into the formula to calculate: ; The results show that the fluctuation frequency correction value is 0.0791, which is consistent with the preset frequency synchronization characteristic judgment interval. The comparison shows that the signal segment falls within the valid range. This result indicates that the current signal segment has dense fluctuations and a synchronous change trend, which meets the criteria for determining a key response node. Therefore, it can continue to be marked as a key response node and included in the trend response number set. Table 2: Time Data Table for Fluctuation Range

[0031] Table 2 lists the original time intervals, offset values, weighting values, and product terms used for formula solving within the fluctuation range of the current signal segment; All parameters are derived from the current signal fluctuation monitoring curve. The time interval is obtained by dividing the intervals using a peak-valley extraction algorithm. The absolute value is obtained by direct difference operation, and the sum of squares is obtained by summing the squares of each term. The dimensionless ratio is normalized by square root to ensure uniform processing of the impact of each time interval on fluctuation assessment. The advantage of the formula is that by introducing the product structure of the absolute value of the relative deviation term and the normalized fluctuation weight term, the fluctuation of each segment is processed in a joint weighted manner, which effectively avoids the problem of overall frequency distortion caused by abnormal local time intervals, and has stronger stability and sensitivity when multiple signal segments are superimposed.

[0032] S204: Based on the device identification code corresponding to the node in the key response node number set of the mutation trend, extract the device number information and reorganize the number. Organize it according to the original index sorting method of the node in the signal sequence, and output the device number of the marked node to obtain the inflection point device node number record. The equipment retrieves the source equipment of each node in the original data one by one. Based on the mapping relationship established between the number and the equipment, the equipment code corresponding to the number is retrieved as the equipment node identification information. The equipment codes are uniformly sorted. The equipment is recombined according to the order in which the number appears in the original signal sequence. Multiple nodes appearing in the same type of equipment are merged to ensure that each equipment identifier is uniquely associated with the response node. The sorted equipment numbers are output in order, reflecting which specific equipment has a significant response change behavior at a certain time period. This forms the basis for subsequent process control and equipment behavior tracking, and the inflection point equipment node number record is obtained.

[0033] Please see Figure 4 The specific steps of S3 are as follows: S301: Based on the equipment number list in the inflection point equipment node number record, classify the equipment types of pick-and-place machine, reflow soldering machine and printing machine respectively, call the signal sequence of each type of equipment in the mutation cycle, retrieve the first fluctuation mutation position in the corresponding signal of the equipment, record the timestamp corresponding to the position, and obtain the equipment first response timestamp group; The system reads the device type information from the complete list of device numbers and categorizes the devices by prefix or database association. Devices consisting of pick-and-place machines, reflow soldering machines, and printing presses are divided into three independent subcategories. Within each category, the system extracts the original signal sequence corresponding to the abrupt change time period, retrieves key fluctuation segments within the target segment, and iterates through the sampling point sequence of each signal category. A sliding difference method is used to detect fluctuation abrupt changes; that is, by calculating the incremental change amplitude between two adjacent signal values, if the increase exceeds a preset fluctuation trigger threshold, a threshold is set to trigger if a continuous stable signal suddenly changes by 5%. If a jump occurs, the point is marked as a fluctuation change point. The timestamp corresponding to that position is extracted and recorded. For example, if the current of a chip mounter suddenly rises from 2.1A to 3.3A at point 380, point 380 can be marked as the first response position of the device. If the corresponding timestamp is "2025-08-14 10:02:15.380", it is recorded in the response timestamp group. This operation will be performed on all numbers in each type of device, and the time point of the first jump will be obtained from its corresponding signal sequence. The time points of the first significant response of different devices will be organized to obtain the device first response timestamp group.

[0034] S302: Based on the timestamps and equipment numbers in the equipment first response timestamp group, prioritize the equipment numbers according to the process sequence of printer, pick and place machine and reflow soldering in the SMT production line. Calculate the difference between the first response timestamp of each equipment and the expected response timestamp based on the process sequence to obtain the equipment response deviation value. Remove equipment numbers with deviation values ​​exceeding a preset threshold. Pair the remaining equipment numbers with their corresponding first response timestamps to obtain a sequence matching response record group. The equipment response deviation value is calculated using the following formula: ; in, The device number is The device response deviation value, The representative equipment number is The corresponding first response timestamp, This represents the average of the first response timestamps. The representative equipment number is Response time weighting factor The device number is The first response timestamp For the device number index variable, Represents the total number of equipment serial numbers; Formula calculation logic: By calculating the deviation of the device's response time from the average response time of all devices, and combining it with the device's weighting factor for normalization, the relative deviation of a single device in the response sequence is quantified. The numerator of the formula is the product of the response offset and the weight, reflecting the speed and importance of the device's response. The denominator is the square root of the sum of squares of the device response time offsets, i.e., the standard deviation, used to unify the dimensions and adjust the scale. By processing the absolute value, the deviation value is made non-negative, which facilitates unified threshold judgment and device screening. The equipment response deviation value represents the standardized distance between the first response time of a certain piece of equipment and the overall average response time. The larger the value, the more abnormal the response of the equipment or the more significant the response delay. Combining the response speed with the importance of the equipment in the process flow, it is an important basis for determining whether the equipment needs to be eliminated. Parameter definition and acquisition instructions: Equipment number is The first response timestamp, in seconds (s), is obtained from the device response record. If device A's response record is set to 2025-08-01 10:25:02 and the start signal is sent at 10:25:00, then its... ; : Average response time of device IDs including response timestamps, based on each The values ​​are taken as the arithmetic mean, and the timestamps of 5 devices (A~E) are set to be... ,but ; Equipment number is The response time weighting factor, used to represent its importance and positional priority, needs to be quantified for non-numerical equipment type data. A process mapping level method is used for conversion: printing machines are set to level 1, placement machines to level 2, and reflow soldering machines to level 3; then standardized to the [0, 1] interval as the weight value: for example, if equipment A is a placement machine, the process level is 2, and the mapping weight is... ; Equipment number is The response timestamp is included in the list of devices with valid timestamps; : An index counter variable for the device number, used to iterate over devices in the summation expression; The total number of device IDs with valid first response timestamps was 5 in the actual data collection. ; Explanation and examples of weight factor settings: Weighting factors The process importance mapping method is adopted, and the specific setting criteria are as follows: Table 3: Weighting Factor Calculation Table

[0035] As shown in Table 3, this standard refers to the typical equipment sequence in the SMT production line and the distribution of its impact on the overall production line delay. For example, the placement machine has a greater impact on the overall production line synchronization due to the equipment response delay, so its weight is set higher than that of the printing machine. Table 4: SMT Equipment Response Time and Weighting Table

[0036] Table 4 lists the actual response data and their weighted conversion values ​​for five devices equipped with response recording devices. Parameter substitution and formula calculation: Taking device A as an example, let's substitute the data from Table 2: ; ; ; ; Substitute into the formula: ; The result indicates that the response deviation of device A is 0.4216. Based on empirical settings, the critical rejection threshold for response deviation is set to 0.8. When a device's response is unstable, it needs to be removed. Device A's deviation value is below the threshold, so it is judged to have a stable response and is retained. This result is used in the device pairing process after removing the item number that does not include the response timestamp. The advantage of the formula lies in the introduction of a device response time weighting factor. Working in conjunction with the standardized deviation structure, it achieves a dual consideration of equipment response timeliness and process criticality; core indicators It can dynamically measure the impact of equipment on overall response stability, providing quantitatively supported reference data in the equipment initial screening and response ranking process.

[0037] S303: Call the combination value of the device number and timestamp in the sequential matching response record group, and concatenate them according to the sorted process order and time sequence to obtain the initial state response timing sequence; The equipment concatenates and integrates the numbers and timestamps according to the established process flow. By traversing the entire record group, it reads the number and response time information one by one. According to the process sequence logic of the printer, pick-and-place machine, and reflow soldering, it sorts the timestamps in each record group in parallel and assembles them into a response process sequence, forming a complete initial state response time sequence. This sequence contains the actual order in which the equipment responds in the actual mutation cycle. The equipment constructs a response chain by reading and connecting the contents of each group of numbers and times. If the printer response time is 10:01:30, the pick-and-place machine is 10:01:35, and the reflow soldering is 10:01:40, the equipment will concatenate the data to form a time sequence linked list. Each element consists of the equipment name and the corresponding response time, and they are arranged in time order as the main line to form a complete structure. This structure is used for time feature comparison and event sequence mapping in subsequent multi-event analysis to obtain the initial state response time sequence.

[0038] Please see Figure 5 The specific steps of S4 are as follows: S401: Call the device number and corresponding timestamp data in the initial state response timing sequence, compare them item by item according to the timestamp from early to late, and determine whether the time sequence between adjacent devices is continuously increasing. If the increasing condition is met, establish a connection relationship between adjacent device numbers, and add them sequentially according to the time order to obtain the first and last connected node path. Based on the timestamp corresponding to each device number, the timestamps are sorted sequentially starting from the earliest time. The timestamp values ​​of any two adjacent devices in the sorted time series are compared to check for a strict increasing trend (i.e., whether the timestamp of the later device is later than the timestamp of the earlier device). If the increasing condition is met, the device confirms a valid response relay relationship between the two devices, immediately establishes a connection between the two numbers, and records them chronologically. During execution, the device uses a loop traversal method, comparing each subsequent node sequentially from the first response node. If a time interval appears at a certain position... If the time of the stamp is equal to or less than that of the previous node, the device skips the pair and does not perform a connection operation. It only retains the device path formed by a strict increasing relationship. If the printer response time is 08:32:15, the pick-and-place machine is 08:32:21, and the reflow soldering is 08:32:25, then these three nodes satisfy the increasing relationship and can be connected into a linear sequence. After each successful connection confirmation, the device appends the current node to the end of the previous connected node, gradually forming a complete chain of nodes connected end to end. This judgment and appending operation is repeated throughout the time series to obtain the chain of nodes connected end to end.

[0039] S402: Based on the sequence of device numbers that have been connected in the first and last serial node path, mark the starting node and the ending node, and mark the propagation position number of each node in the structure from the first node in sequence to obtain the chain control response structure path data. The first device number in the calibration sequence is the starting node, which is the earliest node to trigger in the response time. At the same time, the last device number in the calibration sequence is the ending node, representing the device that responds last in this mutation chain. The devices are labeled in chronological order, starting from the first node and moving backward to mark the propagation position of each device number in the path structure. Starting from the first node, the propagation numbers are assigned sequentially. The first number is labeled as P1, the second number is labeled as P2, and so on until the last node is labeled. Each device number is associated with a unique structural position number, which can trace the relative propagation position of each node in the path and obtain the chain control response structure path data.

[0040] Please see Figure 6 The specific steps of S5 are as follows: S501: Call the first node device number in the chain control response structure path data, obtain the temperature signal sequence and current signal sequence corresponding to the device, extract the continuous rising segment of the temperature signal and the sudden rising segment of the current signal within the mutation cycle, compare them according to the same time axis, and determine whether the two signals appear synchronously in adjacent or overlapping segments. If they appear synchronously, they are marked as overlapping trends, and the mutation mode overlap flag value is obtained. The original temperature and current signal sequences of the device during the mutation cycle are retrieved. The temperature sequence is traversed point by point in the sampling time order. By calculating the difference between adjacent points, continuous rising segments are selected, that is, segments in which the temperature value continues to increase among multiple consecutive sampling points. In the current signal, a sliding differential method is used to search for rapid rise segments in a short period of time, that is, to find time segments in which the current change changes abruptly and continues to exceed a preset amplitude. The start and end time numbers of the temperature rise segment and the current rise segment are recorded respectively. The positions of these two segments are compared on a unified time axis. If the two signals have time overlap or the time number interval does not exceed two to three sampling points, they are identified as adjacent or overlapping segments. The device marks the judgment result as a valid trend overlap relationship. This judgment will be repeatedly performed on multiple similar time periods during the mutation cycle. For example, if the temperature signal continues to rise from point 150 to point 170 and the current signal rises abruptly from point 168 to point 174, there is an intersection of 6 points between the two segments, and the mutation pattern overlap flag value is obtained.

[0041] S502: Based on the initial node device number corresponding to the true value of the mutation mode coincidence flag, extract the waveform of the initial segment of temperature and current signals in the previous cycle, compare it with the waveform corresponding to the node in the mutation interval, and make a synchronization judgment according to the waveform slope direction and change density to obtain the synchronization trend conforming node index set. The equipment needs to trace the complete signal record of the device before the sudden change occurs, and extract the waveform of the initial segment of the temperature and current signals in the previous cycle. This initial segment is generally the complete waveform within the 500ms time period before the sudden change occurs. After extraction, it is compared with the waveform of the same signal in the sudden change interval. During the comparison, the direction of the waveform slope is calculated to determine whether the signal is rising, falling or fluctuating. The change density value is obtained by analyzing the amount of change per unit time. The equipment compares the slopes of the two waveform segments to see if they are in the same direction, i.e. both are rising or falling, and determines whether their change density per unit time falls within the same numerical range. The average increase value of the unit sampling point of the current signal in the previous segment is set to 0.15A, and the sudden change segment is set to 0.14A. If the slope direction of the temperature signal is consistent and the change amplitude is close, it is considered to be in a consistent synchronous trend, and a synchronous trend matching node index set is obtained.

[0042] S503: Call the node number marked in the node index set that matches the synchronization trend, and combine it with the sorting position in the chain control response structure path data to filter the first node that simultaneously meets the characteristics of overlapping mutation mode and trend synchronization. Extract the device identification code of the first node, mark it separately, and obtain the fault tracing device node identification result. Each marked device node number is traversed and matched. Combined with the sorting position information stored in the chain control response structure path data, node numbers that meet two key characteristics are selected. The judgment criteria are that the node has both the mutation mode overlap sign and the trend synchronization conformity sign, and is located at the first node position in the response path. Through the selection process, the first node that constitutes the source response can be selected from multiple candidate device nodes. The device number is extracted separately and mapped to an independent device identifier code. As a device node with obvious linkage source characteristics, it is specially marked. This device identifier is the fault tracing node, indicating that the device is the first to generate the overlapping mutation in the mutation response chain and has the characteristic of consistent trend before and after. The fault tracing device node identifier result is obtained.

[0043] Please see Figure 7 An IoT-based SMT production line equipment fault diagnosis system includes: The signal acquisition module acquires the current signal of the pick-and-place machine, the temperature signal of the reflow soldering machine, and the acceleration signal of the printer. After unifying the sampling period, the timestamps of the signals are normalized to generate identification information of sudden events in the linkage of equipment nodes. Based on the device node linkage sudden event identification information, the trend linkage identification module calls the current signal jump section position and the acceleration signal waveform peak position to cross-locate the jump endpoint and vibration peak, identify whether the two waveforms overlap in time within a continuous period, and generate inflection point device node number record; The initial response sequence generation module collects the first response timestamps of the pick-and-place machine, reflow soldering machine and printing machine during the mutation cycle according to the inflection point equipment node number record, sorts the numbers according to the SMT production line sequence, removes nodes without response data, and generates the initial state response time sequence. The response path construction module determines the connection relationship between the time sequence of SMT production line equipment based on the initial state response timing sequence, constructs a series structure starting from the first node, forms a chain-arranged response path, and obtains the chain-controlled response structure path data. The abnormal node identification module determines the connection relationship between SMT production line equipment based on the initial state response timing sequence, constructs a serial structure starting from the first node, forms a chain-like response path, and obtains the chain control response structure path data.

[0044] The above are merely specific embodiments 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 fault diagnosis of SMT production line equipment based on the Internet of Things, characterized in that, Includes the following steps: S1: Obtain the current signal of the drive motor of the SMT production line pick and place machine, the temperature signal of the reflow soldering heating zone, and the acceleration signal of the printing machine squeegee tip. Perform trend direction judgment on the current jump segment, the temperature continuous rise segment, and the acceleration fluctuation segment, and generate equipment node linkage sudden event identification information. S2: Based on the device node linkage sudden change event identification information, call the current signal jump section position and the acceleration signal waveform peak position, cross-locate the jump end point and vibration peak, identify whether the two waveforms overlap in time within a continuous period, and obtain the inflection point device node number record; S3: Based on the inflection point equipment node number record, collect the first response timestamp of the pick-and-place machine, reflow soldering machine and printing machine during the mutation cycle, sort the numbers according to the SMT production line sequence, remove nodes with no response data, and obtain the initial state response timing sequence. S4: Call the initial state response timing sequence to determine the connection relationship between the time sequence of SMT production line equipment, construct a serial structure starting from the first node, form a chain-like response path, and obtain the chain control response structure path data.

2. The method for fault diagnosis of SMT production line equipment based on the Internet of Things according to claim 1, characterized in that, The device node linkage mutation event identification information includes trend consistency judgment results, signal fluctuation synchronization, and time window alignment information. The inflection point device node number record includes the key response position number of the mutation trend, device identification information, and trend variation point label. The initial state response time sequence includes the response device number, the first response timestamp, and the number sorting result. The chain control response structure path data includes the first node, propagation order, and response path chain arrangement structure.

3. The method for fault diagnosis of SMT production line equipment based on the Internet of Things according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Acquire the current signal of the drive motor of the SMT production line pick and place machine, the temperature signal of the reflow soldering heating zone, and the acceleration signal of the squeegee tip of the printer. Combine the time series of the three types of signals with a unified sampling period to perform synchronization and alignment processing. Call the original timestamp of the signal and match the corresponding sampling point number. Construct a time window based on the time points with the same number to obtain the signal sequence group of the synchronization period. S102: Based on the pick-and-place machine drive current sequence, reflow soldering heating zone temperature sequence, and scraper end acceleration sequence in the synchronous time period signal sequence group, extract the value ranges of the current jump segment, the temperature continuous rise segment, and the acceleration fluctuation segment within the time window, identify the signal amplitude change trend from the start to the end of the interval segment, and obtain the trend consistency direction combination. S103: Based on the trend consistency direction combination, compare the frequency of signal changes in the combination in turn, extract the number of fluctuation peaks and valleys and the interval time interval within the corresponding time window, and determine whether the fluctuation segment fluctuation time sequence overlap condition is met. If the determination result is true, extract the fluctuation overlap interval of the signal combination and mark the time node position to obtain the device node linkage sudden change event identification information.

4. The IoT-based SMT production line equipment fault diagnosis method according to claim 3, characterized in that, The specific steps of S2 are as follows: S201: Based on the device node linkage mutation event identification information, call the position of the current signal in the corresponding jump section and the position of the waveform peak in the acceleration signal, extract the end time point and peak time point of the jump section, compare them according to the same time axis number, and filter whether the two have an intersection in the same time period within the continuous period. If there is an intersection, mark the time point to obtain the jump peak intersection time point set. S202: Based on the time of the intersection of the peak jump time points, extract the position of the intersection point in the original signal sequence, sequentially search forward the starting position of the trend segment of the current signal and acceleration signal, mark the position of the leading node of the intersection point in the continuous trend segment, call the node number for labeling, and obtain the trend segment leading node index group. S203: Call the node number marked in the trend segment leading edge node index group, extract the fluctuation density of current signal and acceleration signal in the preceding and following time periods, calculate the fluctuation frequency correction value, and determine whether the trend dense segment synchronous characteristics are present in the segment. If the frequency range is consistent and the interval difference is within the limited range, then mark the node as a key response point and obtain the set of key response node numbers for sudden trend. S204: Based on the device identifier code corresponding to the node in the key response node number set of the mutation trend, extract the device number information and reorganize the number. Organize the nodes according to the original index sorting method in the signal sequence, and output the device number of the marked node to obtain the inflection point device node number record.

5. The IoT-based SMT production line equipment fault diagnosis method according to claim 4, characterized in that, The fluctuation frequency correction value is a quantized value determined by calculating the first difference or variance of the fluctuation time interval.

6. The IoT-based SMT production line equipment fault diagnosis method according to claim 4, characterized in that, The specific steps for S3 are as follows: S301: Based on the equipment number list in the inflection point equipment node number record, the equipment types of pick-and-place machine, reflow soldering machine and printing machine are classified and processed respectively. The signal sequence of each type of equipment in the mutation cycle is called, the first fluctuation mutation position in the corresponding signal of the equipment is retrieved, and the timestamp corresponding to the position is recorded to obtain the equipment first response timestamp group. S302: Based on the timestamps and equipment numbers in the equipment first response timestamp group, the equipment numbers are prioritized according to the process sequence of printer, pick-and-place machine and reflow soldering in the SMT production line. The equipment response deviation value is obtained by calculating the difference between the first response timestamp of each equipment and the expected response timestamp based on the process sequence. Equipment numbers with deviation values ​​exceeding a preset threshold are removed. The remaining equipment numbers are paired with their corresponding first response timestamps to obtain a sequence matching response record group. S303: Call the combination value of the device number and timestamp in the sequence matching response record group, and splice them according to the sorted process order and time sequence to obtain the initial state response timing sequence.

7. The IoT-based SMT production line equipment fault diagnosis method according to claim 6, characterized in that, The specific steps of S4 are as follows: S401: Call the device number and corresponding timestamp data in the initial state response timing sequence, compare them item by item from early to late according to the timestamp, and determine whether the time sequence between adjacent devices is continuously increasing. If the increasing condition is met, establish a connection relationship between adjacent device numbers, and add them sequentially according to the time order to obtain the first and last connected node path. S402: Based on the sequence of device numbers that have been connected in the first and last connected node paths, the starting node and the ending node are marked, and the propagation position number of each node in the structure is marked sequentially from the first node to obtain the chain control response structure path data.

8. The method for fault diagnosis of SMT production line equipment based on the Internet of Things according to claim 1, characterized in that, The method further includes the following steps: S5: Using the path data of the chain control response structure, determine whether the continuous rise in temperature signal and the sudden rise in current of the first node device have an overlapping trend, analyze whether it has the characteristics of multiple signal types of sudden combination, and identify whether the node is an abnormal initiating node in the chain control structure based on whether the signal starting waveform and the node waveform have a synchronous change trend in the previous cycle, and generate the fault tracing device node identification result. The fault tracing device node identification results include the anomaly initiation node identifier, mutation combination feature matching results, and starting waveform synchronization determination label.

9. The IoT-based SMT production line equipment fault diagnosis method according to claim 8, characterized in that, The specific steps of S5 are as follows: S501: Call the first node device number in the chain control response structure path data, obtain the temperature signal sequence and current signal sequence corresponding to the device, extract the continuous rising segment of the temperature signal and the sudden rising segment of the current signal within the mutation cycle, compare them according to the same time axis, and determine whether the two signals appear synchronously in adjacent or overlapping segments. If they appear synchronously, they are marked as overlapping trends, and the mutation mode overlap flag value is obtained. S502: Based on the initial node device number corresponding to the true value of the mutation mode overlap flag, extract the waveform of the initial segment of temperature and current signals in the previous cycle, compare it with the waveform corresponding to the node in the mutation interval, and make a synchronization judgment according to the waveform slope direction and change density to obtain the synchronization trend conforming node index set. S503: Call the node number marked in the synchronization trend matching node index set, and combine it with the sorting position in the chain control response structure path data to filter the first node that simultaneously meets the characteristics of overlapping mutation mode and trend synchronization. Extract the device identification code of the first node, mark it separately, and obtain the fault tracing device node identification result.

10. A fault diagnosis system for SMT production line equipment based on the Internet of Things, characterized in that, The system is used to implement the IoT-based SMT production line equipment fault diagnosis method according to any one of claims 1-9, the system comprising: The signal acquisition module acquires the current signal of the pick-and-place machine, the temperature signal of the reflow soldering machine, and the acceleration signal of the printer. After unifying the sampling period, the timestamps of the signals are normalized to generate identification information of sudden events in the linkage of equipment nodes. Based on the device node linkage sudden event identification information, the trend linkage identification module calls the current signal jump section position and the acceleration signal waveform peak position to cross-locate the jump endpoint and vibration peak, identify whether the two waveforms overlap in time within a continuous period, and generate inflection point device node number record; The initial response sequence generation module collects the first response timestamps of the pick-and-place machine, reflow soldering machine and printing machine during the mutation cycle according to the inflection point equipment node number record, sorts the numbers according to the SMT production line sequence, removes nodes without response data, and generates the initial state response time sequence. The response path construction module determines the connection relationship between the time sequence of SMT production line equipment based on the initial state response timing sequence, constructs a series structure starting from the first node, forms a chain-arranged response path, and obtains the chain-controlled response structure path data. The abnormal node identification module determines the connection relationship between SMT production line equipment based on the initial state response timing sequence, constructs a serial structure starting from the first node, forms a chain-like response path, and obtains the chain control response structure path data.

Images