Logistics sorting line fatigue early warning method and processing device

Abstract

The application discloses a logistics sorting production line fatigue early warning method and processing equipment, relates to the technical field of logistics sorting production lines, and solves the technical problem of lacking accurate positioning and impact tracing, not considering the working condition difference in fatigue prediction factors, and leading to insufficient structure fatigue prediction accuracy. The method comprises the following steps: receiving an event trigger message uploaded by a trigger node and auxiliary response characteristics uploaded by a cooperative node; determining an impact occurring structure section and actual response characteristics of the impact occurring structure section based on the trigger response characteristics and the auxiliary response characteristics; comparing the actual response characteristics of the impact occurring structure section with baseline response characteristics and historical response characteristics of the impact occurring structure section under a current working condition type and an impact source type, respectively, to obtain baseline deviation and response change; updating the cumulative damage amount of the impact occurring structure section; and obtaining a fatigue grade based on the updated cumulative damage amount, the baseline deviation and the response change.

Description

Technical Field

[0001] This invention relates to the field of logistics sorting production line technology, and in particular to a fatigue early warning method and processing equipment for logistics sorting production lines. Background Technology

[0002] In logistics sorting production lines, structural monitoring nodes are typically deployed at key locations such as the main conveyor belt, narrow-belt sorter and its support, sorting execution units, grids / chutes, and steel structures. Each node integrates an accelerometer and / or strain sensor to detect the vibration, impact, and strain response of the structural section where the node is located. During high-speed operation, logistics sorting production lines continuously endure impact loads from package drops, grid movements, and the start and stop of conveyor sections. Long-term exposure to these loads can lead to fatigue damage in structural components such as supports, beams, roller seats, grid components, and steel structures. Therefore, effective fatigue prediction of structural sections is necessary to prevent production accidents.

[0003] Existing fatigue early warning schemes typically use a fixed signal threshold: when the structural monitoring node detects that the signal amplitude exceeds the preset amplitude threshold, it reports to the host computer, which then determines whether to trigger a fatigue early warning based on the frequency or duration of the over-limit event. However, this method has the following shortcomings: First, logistics sorting production lines have metal structures, and structural responses such as vibration often propagate along the structure. Moreover, the number of nodes is limited, and the structural response signal of a node may originate from the structural response propagation after an impact on an adjacent structural section. Directly regarding the measurement signal of a single node as the impact load response of its structural section will lead to distortion of the impact source location and affect the accuracy of fatigue accumulation. Second, the operating conditions of logistics sorting production lines vary greatly, such as peak and off-peak periods, light and small items and heavy items, continuous sorting operations and single operations, high-speed conveying and low-speed conveying, etc. Different operating conditions cause different impact damage to the structure. Setting a fixed threshold will lead to insufficient accuracy of fatigue warning. For example, it is easy to frequently false alarm during peak periods and may miss early anomalies during off-peak periods. Third, different impact sources (package impact, sorting operation impact, equipment start-up and shutdown impact, etc.) also have different degrees of structural damage.

[0004] In the process of realizing this invention, the inventors discovered that the prior art has at least the following problems: The lack of accurate location and source of impact, as well as the failure to incorporate differences in working conditions into fatigue prediction, leads to insufficient accuracy in structural fatigue prediction. Summary of the Invention

[0005] The purpose of this invention is to provide a fatigue early warning method and processing equipment for logistics sorting production lines, in order to solve the technical problems existing in the prior art, such as the lack of accurate impact location and impact source tracing, and the failure to incorporate differences in working conditions into fatigue prediction, resulting in insufficient accuracy in structural fatigue prediction. The various technical effects of the preferred technical solutions provided by this invention are detailed below.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a fatigue early warning method for a logistics sorting production line. The logistics sorting production line is divided into multiple structural sections, and nodes for structural monitoring are deployed on all or part of these structural sections. The method includes: receiving an event trigger message uploaded by a triggering node and auxiliary response features uploaded by one or more cooperating nodes; the event trigger message includes an impact time window and trigger response features; the structural section where the cooperating node is located is connected to the structural section where the triggering node is located; determining the impact-occurring structural section and its actual response features based on the trigger response features of the triggering node and the auxiliary response features of the cooperating nodes; and determining the impact-occurring structural section and its actual response features based on the impact-occurring... The impact source is determined by the actual response characteristics of the structural segment, the sorting task flow record, and the action flow record. The current working condition type is determined based on the sorting task flow record and the action flow record. The actual response characteristics of the impact-causing structural segment are compared with the baseline response characteristics and historical response characteristics of the impact-causing structural segment under the current working condition type and the impact source type, respectively, to obtain the baseline deviation and response change. The cumulative damage of the impact-causing structural segment is updated based on the actual response characteristics. The fatigue level of the impact-causing structural segment is obtained based on the updated cumulative damage, the baseline deviation, and the response change, and the fatigue level is output.

[0007] Preferably, the method further includes: statistically analyzing the structural sections where the impact occurs before... The fatigue level of the impact-affected structural segment is obtained based on the percentage of peak response exceeding the limit during the second impact, the updated cumulative damage, the baseline deviation, the response change, and the percentage of peak response exceeding the limit. It is a positive integer.

[0008] Preferably, each node is equipped with an acceleration sensor and / or a strain sensor; when the acceleration detected by the acceleration sensor of a node is greater than an acceleration threshold, and / or when the strain detected by the strain sensor of a node is greater than a strain threshold, the node is designated as a trigger node, and the impact time window and trigger response features are extracted from the acceleration time series and / or strain time series.

[0009] Preferably, both the trigger response feature and the auxiliary response feature include the impact arrival time, the response peak value, and the response spectrum distribution vector; the determination of the impact occurrence structural segment and the actual response feature of the impact occurrence structural segment based on the trigger response feature of the trigger node and the auxiliary response feature of the cooperating node includes: comparing the trigger response feature of the trigger node with the auxiliary response feature of all the cooperating nodes to obtain the actual arrival time difference sequence, the actual response peak attenuation ratio sequence, and the actual response spectrum similarity sequence between the trigger node and all the cooperating nodes; and determining multiple candidate structural segments based on the structural topology diagram of the logistics sorting production line. For each candidate structural segment, the following steps are performed: Based on the simulated response database, obtain the theoretical arrival time difference sequence, theoretical response peak attenuation ratio sequence, and theoretical response spectrum similarity sequence of the triggering node and all cooperating nodes when a simulated impact source is applied to the candidate structural segment; based on the similarity between the actual arrival time difference sequence and the theoretical arrival time difference sequence, the similarity between the actual response peak attenuation ratio sequence and the theoretical response peak attenuation ratio sequence, and the similarity between the actual response spectrum similarity sequence and the theoretical response spectrum similarity sequence, obtain the location score of the candidate structural segment; determine the impact-occurring structural segment based on the location scores of all candidate structural segments; and obtain the actual response characteristics of the impact-occurring structural segment based on the actual response characteristics of some nodes on the logistics sorting production line.

[0010] Preferably, the positioning score of each candidate structural segment is a weighted sum of the similarity between the actual arrival time difference sequence and the theoretical arrival time difference sequence, the similarity between the actual response peak attenuation ratio sequence and the theoretical response peak attenuation ratio sequence, and the similarity between the actual response spectrum similarity sequence and the theoretical response spectrum similarity sequence.

[0011] Preferably, determining the impact-causing structural segment based on the location scores of all candidate structural segments includes: sorting the candidate structural segments from high to low according to the location scores; if the difference in location scores between the first-ranked candidate structural segment and the second-ranked candidate structural segment is greater than a preset score difference threshold, then the first-ranked candidate structural segment is taken as the impact-causing structural segment.

[0012] Preferably, determining the impact source type based on the actual response characteristics, sorting task flow records, and action flow records of the impact-occurring structural segment includes: filtering out one or more candidate impact sources occurring in the impact-occurring structural segment from the sorting task flow records and the action flow records; the candidate impact sources include at least one of sorting tasks transporting packages through the impact-occurring structural segment, grid actions in the impact-occurring structural segment, and equipment start-up and shutdown actions; calculating the time matching degree between each candidate impact source and the impact arrival time in the actual response characteristics; calculating the response peak matching degree between each candidate impact source and the response peak value in the actual response characteristics; obtaining the source confidence degree of each candidate impact source based on the time matching degree and the response peak value matching degree; selecting the candidate impact source with the highest source confidence degree as the impact source and determining the impact source type.

[0013] Preferably, updating the cumulative damage of the impact-occurring structural segment based on the actual response characteristics includes: calculating the single-instance basic damage based on the actual response characteristics; correcting the single-instance basic damage using the source confidence of the impact source to obtain the single-instance damage of the impact-occurring structural segment; and adding the single-instance damage to the historical cumulative damage of the impact-occurring structural segment to obtain the updated cumulative damage.

[0014] Preferably, determining the current working condition type based on the sorting task flow record and the action flow record includes: extracting logistics working condition features for a statistical period based on the sorting task flow record and the action flow record, wherein the logistics working condition features include sorting task density, package weight distribution vector, proportion of heavy items, proportion of irregularly shaped items, conveying speed, frequency of compartment actions, and frequency of equipment start-up and shutdown actions; calculating the similarity between the logistics working condition features and the baseline logistics working condition features of different working condition types, and selecting the working condition type with the highest similarity as the current working condition type.

[0015] The present invention also discloses a processing device, comprising: one or more processors; a memory for storing one or more computer programs, wherein the one or more processors are configured to execute the one or more computer programs stored in the memory, so that the one or more processors execute the fatigue early warning method for logistics sorting production lines according to the present invention.

[0016] Implementing one of the above-described technical solutions of the present invention has the following advantages or beneficial effects: This invention enables fatigue prediction of structural sections in logistics sorting production lines. When a trigger event message is received from a trigger node, the auxiliary response characteristics of the trigger node's cooperating nodes are also collected simultaneously. By utilizing the characteristic that the structural responses of the trigger node and its cooperating nodes differ regularly due to their fixed spatial position relationship and vibration propagation path, the impact-causing structural section is located based on the trigger response characteristics and auxiliary response characteristics, and its actual response characteristics are obtained, thus achieving spatial impact positioning. After identifying the structural segment where the impact occurred, this invention jointly determines the impact source based on the actual response characteristics of the structural segment, sorting task flow records, and action flow records. It then determines the current operating condition type based on the sorting task flow records and action flow records. The actual response characteristics of the structural segment where the impact occurred are compared in two dimensions with the baseline response characteristics and historical response characteristics of the same structural segment under the same operating condition type and impact source type. This yields the baseline deviation reflecting the current degree of anomaly and the response change reflecting the direction of structural response change. This avoids the problem of traditional fixed thresholds misjudging normal high loads during peak periods as abnormal or underreporting early degradation under low loads. This invention further updates the cumulative damage of the impacted structural segment by the actual response characteristics. The fatigue level is obtained by fusing and mapping the updated cumulative damage, baseline deviation, and response change from three dimensions. This allows the fatigue level to take into account the long-term life consumption (cumulative damage), the current degree of abnormality (baseline deviation), and the direction of structural response change (response change), thereby improving the accuracy and engineering practicality of structural fatigue early warning in logistics sorting production lines. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of 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. In the drawings: Figure 1 This is a flowchart illustrating the fatigue early warning method for logistics sorting production lines in Embodiment 1 of the present invention. Figure 2 This is an internal connection block diagram of a node in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the transmission between the triggering node and the cooperating node when an impact event occurs in Embodiment 1 of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, various exemplary embodiments described below will be referenced to the accompanying drawings, which form part of the exemplary embodiments, illustrating various exemplary embodiments that may be used to implement the present invention. Unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. It should be understood that they are merely examples of processes, methods, and apparatuses consistent with some aspects of the present invention disclosed as detailed in the appended claims, and other embodiments may be used, or structural and functional modifications may be made to the embodiments listed herein without departing from the scope and spirit of the present invention.

[0019] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," etc., indicate the orientation or positional relationship based on the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the referred element must have a specific orientation, or be constructed and operated in a specific orientation. The terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. The term "multiple" means two or more. The terms "connected" and "linked" should be interpreted broadly, for example, they can be fixed connections, detachable connections, integral connections, mechanical connections, electrical connections, communication connections, direct connections, indirect connections through an intermediate medium, and can be the internal connection of two elements or the interaction relationship between two elements. The term "and / or" includes any and all combinations of one or more of the related listed items. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0020] To illustrate the technical solution described in this invention, specific embodiments are described below, showing only the parts related to the embodiments of this invention.

[0021] Example 1: This invention discloses a fatigue early warning method for a logistics sorting production line. In this embodiment, the logistics sorting production line is divided into multiple structural sections, and nodes for structural monitoring are deployed on all or part of the structural sections.

[0022] A typical logistics sorting production line generally includes a main conveyor belt, a narrow-belt sorter and its support structure (containing electric rollers), sorting slots / chutes, sorting slot supports, branch conveyor lines, and the steel structure supporting the above equipment or mechanisms. The narrow-belt sorter and its support structure includes roller seats, electric rollers on the roller seats, and the drive assembly for the electric rollers. The sorting slot supports are used to connect the diversion gates, swing wheels, flaps, or chutes to the steel structure.

[0023] Based on the structural design diagram of the logistics sorting production line, the line is divided into multiple structural sections, such as the sorting area, multiple structural sections on the main conveyor belt, one or more structural sections on the narrow belt sorter, and one or more structural sections on the branch conveyor line. A structural section refers to a structural unit in the logistics sorting production line that is relatively independent in mechanical structure and has identifiable boundaries along the vibration / impact propagation path. Specific division rules can be as follows: Structural sections are divided based on the integrity of the vibration propagation path. Specifically, structural components with continuous vibration propagation paths and consistent signal transmission characteristics are divided into the same structural section, using the connection surface of the structural components as the boundary. Structural components whose signal transmission characteristics change significantly after crossing the connection surface are divided into different structural sections. Alternatively, the line can be divided into structural sections of different lengths along the transport path of the sorted packages.

[0024] In this embodiment, the structural segments are numbered and used as graph nodes. At the same time, the entrance and each compartment of the logistics sorting production line are also set as graph nodes. An edge is set between two graph nodes that are actually connected to form a structural topology graph of the logistics sorting production line.

[0025] In this embodiment, nodes for structural monitoring are deployed in all or part of the structural sections. Preferably, the nodes are deployed in structural sections subjected to heavy impact loads, such as grid structural sections, structural sections where the roller seats in the narrow-strip sorting machine and support are located, structural sections at the packaging entrance on branch conveyors, and structural sections where drive equipment starts and stops. Figure 2 As shown, the node includes an embedded processor (such as a microcontroller), a sensing and communication unit, and one or more structural monitoring sensors. The embedded processor is connected and communicates with the sensing and communication unit and the structural monitoring sensors, respectively. The one or more structural monitoring sensors include accelerometers and / or strain sensors. The accelerometer is not limited to a triaxial MEMS (Micro-Electro-Mechanical Systems) accelerometer or a piezoelectric accelerometer. The strain sensor is not limited to a resistance strain gauge or a fiber Bragg grating strain sensor.

[0026] The fatigue early warning method for logistics sorting production lines disclosed in this embodiment is not limited to the host computer of the logistics sorting production line, or other computer equipment or servers. For example... Figure 1 As shown, the method includes: Step S1: Receive the event trigger message uploaded by the triggering node and the auxiliary response features uploaded by one or more cooperating nodes. The event trigger message includes the impact time window and the trigger response features; such as... Figure 3 As shown, the structural segment where the collaborative node is located is connected to the structural segment where the trigger node is located.

[0027] In this embodiment, all nodes on the logistics sorting production line utilize structural monitoring sensors to collect the initial structural response signal sequence (including initial acceleration time sequence, initial strain time sequence, etc.) of their respective structural sections at a set sampling rate. The initial structural response signals undergo preprocessing such as DC removal, outlier elimination, and bandpass filtering to obtain structural response signal sequences (e.g., acceleration time sequence, strain time sequence, etc.) to reduce noise interference. Nodes can use a circular buffer to store the structural response signal sequence. During the operation of the logistics sorting production line, all nodes synchronously determine impact events. Specifically, they compare the currently collected structural response signal (which may be the signal amplitude or the average signal value within a short period) with a signal threshold. If the currently collected structural response signal is greater than the signal threshold, an impact event is considered to have occurred, and the node that detected the impact event becomes the trigger node. The signal threshold is greater than the background noise and can be set empirically.

[0028] Each node is equipped with an accelerometer and / or a strain sensor. When the acceleration (amplitude or short-period average of the acceleration time series) detected by the node's accelerometer exceeds an acceleration threshold, and / or when the strain (amplitude or short-period average of the strain time series) detected by the node's strain sensor exceeds a strain threshold, that node is designated as a trigger node. It should be noted that due to the complex operating conditions of the logistics sorting production line, there may be more than one trigger node at the same time.

[0029] In this embodiment, the triggering node determines the impact time window based on the structural response signal sequence in its own circular buffer, such as... Figure 3 As shown, the impact event is reported to the gateway of the logistics sorting production line system. The impact event includes an impact time window. Subsequently, the triggering node extracts the trigger response features based on the structural response signal sequence in its own circular buffer, generates an event trigger message, and reports it to the execution entity of this method (such as the host computer of the logistics sorting production line) through the gateway. The gateway determines all structural segments connected to the structural segment where the triggering node is located based on the structural topology diagram of the logistics sorting production line. The nodes deployed on the determined structural segments are designated as collaborative nodes, and a freeze command (containing the impact time window) is issued to all collaborative nodes. The collaborative nodes freeze their own circular buffers according to the freeze command, and extract auxiliary response features from the structural response signal sequence they have collected based on the impact time window. The auxiliary response features are reported to the execution entity of this method (such as the host computer of the logistics sorting production line) through the gateway, and the circular buffers are unfrozen. Freezing prevents relevant data from overflowing from the circular buffers.

[0030] For example, the node index of the trigger node is , The trigger node is an integer; the method for determining the impact time window is as follows: when an impact event is detected, the trigger node is calculated. The envelopes of all obtained structural response signal sequences are used to backtrack from the moment the impact event is determined, and the most recent time point on each envelope whose amplitude does not exceed the background noise threshold (the sum of the background mean and three times the background standard deviation) is selected as the candidate arrival time point. The earliest arrival candidate time point among all arrival candidate time points of all envelopes (such as acceleration envelope and strain envelope) is taken as the impact arrival time. Starting from the moment the impact arrives, the last time point in each envelope that exceeds the background noise threshold is searched backwards as the candidate time point for termination. The latest candidate time point among all the candidate time points for termination of the envelopes is taken as the moment the impact ends. time interval As a window of opportunity for impact. This represents the time offset, and its value is not limited to 0.1s to 1s, but can be applied to any time interval. The duration of the trigger node's response.

[0031] In this embodiment, the trigger node extracts structural response signal sequence segments (including acceleration sequence segments and strain sequence segments) corresponding to the response duration from its own circular buffer, and calculates the response peak value, impact energy, strain cycle amplitude, and response spectrum distribution vector of the structural response signal sequence segments. The response peak value is the amplitude value with the largest absolute value among all acceleration sequence segments. The impact energy is the sum of the squares of the amplitudes at all times in the acceleration sequence segment. The strain cycle amplitude is obtained by processing the strain sequence segment using the rainflow counting method. A fast Fourier transform is performed on the acceleration sequence segment, dividing the spectrum into several continuous frequency bands. The sum of the squares of the spectral amplitudes or the mean energy value within each frequency band is calculated and used as the energy value for that band. The energy values ​​of all frequency bands are arranged into a vector, which serves as the response spectrum distribution vector. The cooperating nodes of the trigger node extract structural response signal sequence segments from their own structural response signal sequences according to the impact time window. Referring to the processing method of the trigger node, they can extract the impact arrival time, response duration, response peak value, impact energy, strain cycle amplitude, and response spectrum distribution vector from the structural response signal sequence segments.

[0032] In this embodiment, the triggering node and the cooperating node use all or part of the impact arrival time, response duration, response peak value, impact energy, strain cycle amplitude, and response spectrum distribution vector they calculate as response features, which are respectively denoted as triggering response features and cooperating response features. To improve the accuracy of fatigue prediction, the execution subject, nodes, and gateway of this method all use the same time base, aligning the time of the triggering response features and the auxiliary response features.

[0033] Step S2 involves determining the impact-generating structural segment and its actual response characteristics based on the triggering response characteristics of the triggering node and the auxiliary response characteristics of the cooperating node. Preferably, both the triggering response characteristics and the auxiliary response characteristics include the impact arrival time, the response peak value, and the response spectrum distribution vector. Step S2 includes: Step S21: Compare the trigger response characteristics of the trigger node with the auxiliary response characteristics of all cooperating nodes to obtain the actual arrival time difference sequence, the actual response peak attenuation ratio sequence, and the actual response spectrum similarity sequence between the trigger node and all cooperating nodes.

[0034] Specifically, trigger node have One collaborative node, A positive integer; trigger node The trigger response characteristics are: , Indicates the trigger node peak response Indicates the trigger node The response spectrum distribution vector; cooperative nodes The auxiliary response characteristics are: , , Represents a collaborative node The moment the impact arrives, Represents a collaborative node peak response Represents a collaborative node The response spectrum distribution vector.

[0035] Calculate and obtain the trigger node and The actual arrival time difference sequence of the coordinating nodes is as follows: .

[0036] Calculate and obtain the trigger node and The actual peak attenuation ratio sequence of each coordinating node is as follows: , Let be the first constant, and its range of values ​​can be . .

[0037] Calculate and obtain the trigger node and The actual response spectrum similarity sequence of the collaborative nodes is as follows: , This represents a similarity calculation function, but is not limited to a cosine similarity calculation function.

[0038] Step S22: Determine multiple candidate structural segments based on the structural topology diagram of the logistics sorting production line. There are M candidate structural segments, where M is a positive integer and m is the index of the candidate structural segment.

[0039] Specifically, from the structural topology diagram, the structural segments that are connected to the trigger node, the structural segments that are not connected to the trigger node but whose vibrations can propagate to the trigger node (i.e., there is a vibration propagation path between the structural region and the trigger node) are all selected as candidate structural segments.

[0040] Step S23: For each candidate structural segment, perform the following: Based on the simulation response database, obtain the theoretical arrival time difference sequence, theoretical response peak attenuation ratio sequence, and theoretical response spectrum similarity sequence of the triggering node and all cooperating nodes when a simulated impact source is applied to the candidate structural segment; Based on the similarity between the actual arrival time difference sequence and the theoretical arrival time difference sequence, the similarity between the actual response peak attenuation ratio sequence and the theoretical response peak attenuation ratio sequence, and the similarity between the actual response spectrum similarity sequence and the theoretical response spectrum similarity sequence, obtain the localization score of the candidate structural segment.

[0041] Specifically, a simulated response database is pre-built, which includes execution within explicit dynamic analysis software: Step 1: Based on the structural design diagram of the logistics sorting production line, construct a digital twin model of the logistics sorting production line in the simulation system, and divide the digital twin model into structural segments according to the structural topology diagram. Step 2: Apply a simulated impact source to each structural segment of the digital twin model and obtain the theoretical response characteristics of all structural segments at this time. The theoretical response characteristics include all or part of the impact arrival time, response duration, response peak value, impact energy, strain cycle amplitude, and response spectrum distribution vector. The structural segment to which the simulated impact source is applied is associated with the theoretical response characteristics of all structural segments and stored in the simulated response database. The simulated impact source is used to simulate the impact load of packages falling into the structural segment during sorting tasks, the impact load caused by the opening action of the structural segment, or the impact load caused by the start-up and shutdown actions of equipment on the structural segment (mainly referring to the start-up and shutdown of rollers). The simulated response database can be deployed in the memory of the execution entity of this method, or it can be deployed on a backend server connected and communicating with the execution entity.

[0042] In this embodiment, the explicit dynamic analysis software is not limited to Abaqus / Explicit, LS-DYNA, or Simcenter Radioss.

[0043] Specifically, For the candidate structure segment index, in step S23, for the candidate structure segment... The execution process is as follows: Step S231: Read the candidate structure segment from the simulation response database. When a simulated impact source is applied, the trigger node is activated. and Based on the theoretical response characteristics of each coordinating node, and referring to the calculation method in step S21 above, the triggering node is obtained. and The following sequence of cooperating nodes: Theoretical arrival time difference sequence: ; Theoretical peak-to-peak decay ratio sequence: ; Theoretical response spectrum similarity sequence: .

[0044] in, Indicates the trigger node The theoretical impact arrives at the moment, Indicates the trigger node The theoretical impact time of the first cooperating node, This represents the theoretical arrival time of the impact from the second collaborating node of triggering node i. Indicates the trigger node Collaborating nodes The theoretical impact arrived at the moment. Indicates the trigger node The theoretical peak response, Indicates the trigger node The theoretical peak response of the first cooperating node, Indicates the trigger node Collaborating nodes The theoretical peak response, Indicates the trigger node Collaborating nodes The theoretical peak response, Indicates the trigger node The theoretical response spectrum distribution vector, Indicates the trigger node The theoretical response spectrum distribution vector of the first cooperating node. Indicates the trigger node Collaborating nodes The theoretical response spectrum distribution vector, Indicates the trigger node Collaborating nodes The theoretical response spectrum distribution vector.

[0045] Step S232: Calculate the similarity between the actual arrival time difference sequence and the theoretical arrival time difference sequence, and denote it as the first similarity. Calculate the similarity between the actual response peak decay ratio sequence and the theoretical response peak decay ratio sequence, denoted as the second similarity. Calculate the similarity between the actual response spectrum similarity sequence and the theoretical response spectrum similarity sequence, denoted as the third similarity. Weighted summation of first similarity Second similarity and third similarity Obtain candidate structural segments Positioning score , , , , These represent the first, second, and third scoring weights, respectively, with values ​​ranging from [value range missing]. One can choose based on experience, which satisfies... Using a weighted sum of first, second, and third similarity scores to obtain the localization score of candidate structural segments can enhance the robustness and accuracy of localization.

[0046] Step S23 involves pre-constructing a simulated response database containing the theoretical response characteristics of all structural segments using explicit dynamic analysis software. This database is directly read during the execution of this invention, improving execution efficiency. Since the structure of the logistics sorting production line is fixed, when an impact event occurs in any structural segment, the vibration propagation path between the triggering node and all its cooperating nodes and that structural segment is fixed. Therefore, when an impact event occurs in each structural segment, regardless of the impact source type, the distribution of the arrival time difference sequence, response peak attenuation ratio sequence, and response spectrum similarity sequence between the triggering node and all cooperating nodes follows a certain pattern. Step S23, based on this pattern, determines the location score of candidate structural segments according to the similarity between the actual arrival time difference sequence and the theoretical arrival time difference sequence, the similarity between the actual response peak attenuation ratio sequence and the theoretical response peak attenuation ratio sequence, and the similarity between the actual response spectrum similarity sequence and the theoretical response spectrum similarity sequence. This allows for the location of the impacted structural segment based on the location score, demonstrating good location robustness and accuracy.

[0047] Step S24: Determine the impact-occurring structural segment based on the location scores of all candidate structural segments.

[0048] Specifically, in The candidate structural segment with the highest positioning score among the candidate structural segments is selected as the impact occurrence structural segment.

[0049] In a preferred embodiment of this example, candidate structural segments are sorted from highest to lowest based on their location scores. If the difference in location scores between the first-ranked candidate structural segment and the second-ranked candidate structural segment is greater than a preset score difference threshold, then the first-ranked candidate structural segment is selected as the impact-occurring structural segment. The preset score difference threshold can be determined based on the location score of the first-ranked candidate structural segment, and its range is set to 15% to 30% of the location score of the first-ranked candidate structural segment to adapt to changes in the location score. When the difference in location scores between the first-ranked candidate structural segment and the second-ranked candidate structural segment is not greater than the preset score difference threshold, the current impact event processing can be terminated. Alternatively, the top three candidate structural segments can be selected as impact-occurring structural segments, and step S3 can be executed for each impact-occurring structural segment to obtain the source confidence of the impact source for each impact-occurring structural segment. The impact-occurring structural segment with the highest source confidence is selected as the final structural segment.

[0050] This preferred embodiment first selects the candidate structural segment with a significantly higher positioning score than other candidate structural segments as the impact occurrence structural segment. This reduces the probability of positioning errors when the positioning scores of multiple candidate structural segments are close due to simulation errors, sensor noise, or production line structural symmetry. When the difference in positioning scores is not significant, the final impact occurrence structural segment is determined by combining the source confidence in the subsequent step S3. This can reduce the risk of subsequent fatigue prediction errors caused by positioning errors.

[0051] Step S25: Obtain the actual response characteristics of the impact-causing structural section based on the actual response characteristics of some nodes on the logistics sorting production line. Preferably, if nodes are deployed in the impact-causing structural section, the actual response characteristics of the deployed nodes are used as the actual response characteristics of the impact-causing structural section. If no nodes are deployed in the impact-causing structural section, the actual response characteristics of the impact-causing structural section are derived based on the trigger response characteristics of the triggering nodes and / or the auxiliary response characteristics of the cooperating nodes.

[0052] It should be noted that the actual response characteristics of the nodes deployed in the impact-affected structural section can be extracted using the method described above for extracting auxiliary response characteristics, and will not be repeated here. The actual response characteristics of the impact-affected structural section include all or part of the following: impact arrival time, response duration, response peak value, impact energy, strain cycle amplitude, and response spectrum distribution vector.

[0053] When it is necessary to deduce the actual response characteristics of the impact-generating structural segment based on the trigger response characteristics of the triggering node and / or the auxiliary response characteristics of the cooperating node, it is assumed that the index of the impact-generating structural segment is... Trigger node and its corresponding Composed of several collaborative nodes Select from the inversion nodes. One inversion node, ,based on The actual response characteristics of each inversion node (when the inversion node is a trigger node, its actual response characteristics are the trigger response characteristics; when the inversion node is a cooperative node, its actual response characteristics are the auxiliary response characteristics) and the actual response characteristics of the impact-affected structural segment derived from the structural attenuation model. The specific process includes: Step S251: Obtain the inversion node of the impact-occurring structural segment (specifically its center point). The distance of the propagation path , ; Step S252, using a pre-calibrated attenuation model based on the inversion nodes peak response Strain Cycle Amplitude Impact energy Response spectrum distribution vector and distance of propagation path Calculate the inversion nodes The derived peak response of the impact-generating structural segment Strain Cycle Amplitude Impact energy and response spectrum distribution vector : ; ; ; ; in, , , , These represent the first, second, third, and fourth attenuation coefficients, respectively. They can be measured by manually tapping at different points on the material sorting production line to determine the response characteristics and perform fitting and calibration. , , , . Represents the natural constant. Indicates the inversion node response spectrum distribution vector Mid-frequency band Response energy, Indicates the inversion node The derived response spectrum distribution vector of the impact-generating structural segment Mid-frequency band The response energy components. Response energy components across all frequency bands. Forming inversion nodes The derived response spectrum distribution vector of the impact-generating structural segment .

[0054] Step S253, Integration Peak response of the impact-affected structural segment derived from each inversion node. Strain Cycle Amplitude Impact energy and response spectrum distribution vector To obtain the peak response of the structural segment where the impact occurred. Strain Cycle Amplitude Impact energy and response spectrum distribution vector The integration method is not limited to weighted average or median.

[0055] Step S254: The impact arrival time and response duration of the inversion node with the smallest propagation path distance to the impact-occurring structural segment are used as the impact arrival time and response duration of the impact-occurring structural segment, with minimal difference between them. The impact arrival time, response duration, and response peak value are then used as the reference values. Strain Cycle Amplitude Impact energy and response spectrum distribution vector The actual response characteristics of the structural segments that are all or part of the impact-affected area.

[0056] Step S25 utilizes respectively The actual response characteristics of each inversion node are used to deduce the actual response characteristics of the impact-affected structural segment. Multi-source fusion is then performed using the weighted average method or the median method, which effectively eliminates the deviation in actual response characteristics caused by differences in the installation location, local noise, or coupling differences of a single node.

[0057] Step S3: Determine the source of the impact based on the actual response characteristics of the impact-causing structural section, the sorting task flow record, and the action flow record; determine the current working condition type based on the sorting task flow record and the action flow record.

[0058] In this embodiment, the executing entity of this method collects sorting task flows in real time from devices or systems such as the warehouse management system, barcode scanners on the production line, weighing equipment, volume measuring equipment, and photoelectric detectors on the logistics sorting production line, forming sorting task flow records. Each sorting task flow includes at least package identification, package attributes, target compartment, planned start time, and sorting path. Package attributes include weight, dimensions, and volume. The executing entity of this method also collects motion flows that can cause fatigue damage to the structure from the material sorting production line in real time, forming motion flow records. Motion flows include at least conveyor speed, compartment action time, compartment action completion status, roller start / stop time, and roller start or stop completion status.

[0059] In a preferred embodiment of this example, step S3, determining the source of the impact based on the actual response characteristics of the impact-causing structural segment, the sorting task flow record, and the action flow record, includes: Step S3a1: Select one or more candidate impact sources from the sorting task flow record and action flow record. The candidate impact sources include at least one of the following: sorting tasks that transport packages through the impact-causing structural section, grid actions performed in the impact-causing structural section, and equipment start-stop actions.

[0060] Step S3a2: Calculate the time matching degree between each candidate impact source and the impact arrival time in the actual response characteristics.

[0061] Specifically, if the candidate impact source is a sorting task, the time when the package corresponding to the sorting task falls into the impact-causing structural segment is determined based on the planned start time and sorting path in the corresponding sorting task flow. The absolute value of the time difference between the package's fall time and the impact arrival time in the actual response characteristics of the impact-causing structural segment is calculated. This absolute value is then normalized to obtain the time matching degree between the candidate impact source and the impact arrival time in the actual response characteristics.

[0062] If the candidate impact source is a grid action or equipment start-up / stop action, then extract the grid action time from the corresponding action flow, or the roller start-up / stop time, and record it as the action time. Calculate the absolute value of the time difference between the action time and the impact arrival time in the actual response characteristics of the impact-occurring structural segment. Normalize the absolute value of the time difference to obtain the time matching degree between the candidate impact source and the impact arrival time in the actual response characteristics.

[0063] Step S3a3: Calculate the response peak matching degree between each candidate impact source and the response peak in the actual response characteristics.

[0064] Specifically, the peak response range differs depending on the type of shock source.

[0065] If the candidate impact source is a sorting task, the weight class of the package is determined according to the package attributes in the corresponding sorting task flow. Different weight classes correspond to different package response peak ranges. If the response peak in the actual response characteristics of the impact occurrence structural segment falls into the package response peak range corresponding to the weight class of the sorting task package, the response peak matching degree is set to 1. Otherwise, the distance between the response peak of the actual response characteristics and the package response peak range corresponding to the weight class of the sorting task package is calculated, and the distance is normalized to obtain the response peak matching degree.

[0066] If the candidate impact source is a grid action or a device start-stop action, when the peak value of the actual response characteristic of the impact-occurring structural segment falls into the corresponding action response peak range (grid action response peak range or device start-stop action response peak range), the response peak matching degree is set to 1. Otherwise, the distance between the peak value of the actual response characteristic and the corresponding action response peak range is calculated, and the distance is normalized to obtain the response peak matching degree.

[0067] Step S3a4: Obtain the source confidence of each candidate impact source based on the time matching degree and the response peak matching degree.

[0068] Specifically, the source confidence of each candidate shock source is obtained by weighted summing of the time matching degree and the peak response matching degree. The weights for the weighted summation can be set by the user based on experience.

[0069] Step S3a5: Select the candidate impact source with the highest source confidence as the impact source and determine the impact source type. Impact source types include sorting tasks, grid actions, and equipment start-up and shutdown actions.

[0070] This preferred embodiment first filters out multiple candidate impact sources by the spatial constraints of the impact occurrence zone, and then calculates the source confidence of the candidate impact sources by combining the time matching degree and the response peak matching degree, which can improve the accuracy of impact source location.

[0071] In a preferred embodiment of this example, step S3, determining the current working condition type based on the sorting task flow record and action flow record, includes: Step S3b1: Extract the logistics condition characteristics of the statistical period based on the sorting task flow record and action flow record. The logistics condition characteristics include sorting task density, package weight distribution vector, proportion of heavy items, proportion of irregularly shaped items, conveying speed, frequency of compartment actions, and frequency of equipment start-up and shutdown actions.

[0072] Specifically, the statistical period is a time interval with the current time as the endpoint and a preset time length in the past. The preset time length is not limited to 30 minutes or 1 hour. Sorting task density represents the number of sorting tasks executed by the logistics sorting line per unit time (e.g., 1 minute). The method for obtaining the package weight distribution vector is as follows: Multiple weight ranges are preset (e.g., 0 to 1kg, 1kg to 5kg, 5kg to 10kg, 10kg to 20kg, and above 20kg). The number of packages falling into each weight range within the statistical period is counted. These package numbers are divided by the total number of packages processed by the logistics sorting line within the statistical period to obtain the proportion of each weight range. The proportions of multiple weight ranges form the package weight distribution vector. The ratio of packages exceeding a preset weight threshold (e.g., 10kg) to the total number of packages within the statistical period is calculated; this ratio is the heavy item proportion. Irregularly shaped packages are those exceeding the standard size range, such as extra-long or extra-tall packages. These irregularly shaped packages are prone to jamming on the production line, affecting conveying speed. The irregularly shaped package proportion is obtained by calculating the ratio of the number of irregularly shaped packages to the total number of packages within the statistical period. The frequency of grid cell operation and the frequency of equipment start-stop operation represent the number of grid cell operations and the number of equipment start-stop operations per unit time (e.g., 1 minute), respectively.

[0073] Step S3b2: Calculate the similarity between the logistics condition characteristics and the baseline logistics condition characteristics of different condition types, and select the condition type with the highest similarity as the current condition type.

[0074] Specifically, the executing entity pre-clusters the logistics condition characteristics within multiple historical statistical periods to obtain several clusters. Each cluster corresponds to a specific condition type, and the cluster center of each cluster serves as the baseline logistics condition characteristic for its corresponding condition type. The baseline logistics condition characteristics for each condition type are pre-stored in memory, and are not limited to the similarity calculated in step S3b2 using cosine similarity.

[0075] This preferred embodiment addresses the traditional fixed-threshold fatigue assessment scheme by constructing logistics condition characteristics from multiple dimensions, such as sorting task density, package weight distribution vector, proportion of heavy items, proportion of irregularly shaped items, conveying speed, frequency of compartment actions, and frequency of equipment start-up and shutdown actions. The operating condition type is determined based on the similarity between the logistics condition characteristics and the baseline logistics condition characteristics of different operating condition types, so that the baseline deviation can be accurately evaluated for different operating condition types, thereby improving the accuracy of fatigue prediction.

[0076] Step S4: Compare the actual response characteristics of the impact-causing structural section with the baseline response characteristics and historical response characteristics of the impact-causing structural section under the current working condition type and impact source type, respectively, to obtain the baseline deviation and response change.

[0077] In this embodiment, the number of feature dimensions of the stored baseline response features and historical response features is less than the number of feature dimensions of the actual response features. The baseline response features and historical response features do not include the impact arrival time. The historical response features are the actual response features that do not include the impact arrival time, calculated according to the aforementioned steps S1, S2, and S3 when the impact-affected structural segment is subjected to an impact event of the same working condition and the same type of impact source.

[0078] Baseline response characteristics of each structural section under different operating conditions and impact sources are pre-established and stored. Specifically, structural response signal sequences obtained by all nodes under different operating conditions and impact sources are collected when the logistics sorting production line is operating normally and the equipment on the production line (such as structural monitoring sensors, grid execution units, etc.) is in normal condition. Based on the structural response signal sequences of each node, the actual response characteristics of that node are extracted. The extraction method can refer to the aforementioned auxiliary response feature extraction process, which will not be repeated here.

[0079] For structural sections with deployed nodes, the actual response characteristics of the nodes are used as the actual response characteristics of the structural section. For structural sections without deployed nodes, their actual response characteristics can be derived using the actual response characteristics of one or more surrounding nodes. The specific derivation process can be referred to in steps S251 to S254, and will not be repeated here. The actual response characteristics of each structural section subjected to two or more impacts under the same working condition and impact source type are integrated into the baseline response characteristics of the structure under that working condition and impact source type. The integration method is not limited to averaging, taking the median, or a combination thereof.

[0080] The memory is associated with storage condition type, impact source type, structural segment and baseline response characteristics, as well as historical response characteristics.

[0081] In this embodiment, step S4 includes: comparing the actual response characteristics of the impact-occurring structural segment with the baseline response characteristics of the impact-occurring structural segment under the current working condition and impact source type to obtain the baseline deviation. It should be noted that only the common feature dimensions of the actual response characteristics and the baseline response characteristics are compared. Specifically, there are four numerical features: response duration, response peak value, impact energy, and strain cycle amplitude. For each numerical feature, the difference between the numerical feature in the actual response characteristics and the corresponding numerical feature in the baseline response characteristics is calculated, and this difference is divided by the corresponding numerical feature in the baseline response characteristics to obtain the deviation of this numerical feature. The KL divergence between the response spectrum distribution vector in the actual response characteristics and the response spectrum distribution vector in the baseline response characteristics is calculated, and the KL divergence is standardized to obtain the deviation of the response spectrum distribution vector. The deviations of all feature dimensions involved in the comparison are weighted and summed to obtain the baseline deviation. The weights of each feature dimension in the weighted summation can be set empirically.

[0082] In this embodiment, step S4 further includes: comparing the actual response characteristics of the impact-generating structural segment with the historical response characteristics of the impact-generating structural segment under the current working condition type and impact source type to obtain the response change.

[0083] It should be noted that only the common numerical dimensions of the actual and historical response characteristics are compared. The difference between the numerical characteristics in the actual response and the corresponding numerical characteristics in the historical response is calculated. This difference is then divided by the corresponding numerical characteristic in the historical response to obtain the trend of change of this numerical characteristic. The numerical characteristic can be the response duration, peak response, impact energy, or strain cycle amplitude. The trends of change of all the compared feature dimensions are then summed using weighted averages to obtain the response change. The weights of each feature dimension in the weighted summation can be set empirically.

[0084] Step S4 departs from the traditional fixed threshold model, employing a comparison between baseline response characteristics and actual response characteristics, subdivided into three dimensions: structural segment, operating condition type, and impact source type. This avoids misjudgments arising from comparisons under different operating conditions and impact source types. The baseline deviation characterizes the degree of deviation from the baseline. Simultaneously, by comparing historical response characteristics with actual response characteristics, the change in response is obtained, indicating the direction of structural response evolution. Using both baseline deviation and response change as indicators of structural fatigue helps improve the accuracy of fatigue prediction.

[0085] Step S5: Update the cumulative damage amount of the impact-affected structural segment based on the actual response characteristics. Preferably, step S5 includes: Step S51: Calculate the basic damage for a single event based on the actual response characteristics.

[0086] Specifically, if the actual response characteristics of the structural segment where the impact occurred include the strain cycle amplitude (expressed as...) Then, based on the strain cycle amplitude, query the SN curve (stress-life curve) of the material in the impact-affected structural section or - The strain-life curve yields the number of fatigue failure cycles corresponding to the strain cycle amplitude. Then the basic damage in a single instance is: .

[0087] If the actual response characteristics of the structural segment where the impact occurred do not include the strain cycle amplitude, but include the impact energy and response duration Then the basic damage in a single instance is: ; in, , The impact energy and response duration represent the baseline response characteristics of the impact-affected structural section under the current operating condition and the impact source type, respectively. Indicates the first experience index, ; This represents the second experience index. . , The value of can be determined experimentally. A single basic damage is a dimensionless scalar.

[0088] Step S52: Correct the single-event basic damage using the source confidence of the impact source to obtain the single-event damage amount of the structural segment where the impact occurred: ; in, The source confidence level of the impact source is calculated through step S3a4 above. By correcting the single basic damage using the source confidence level of the impact source, the contribution of less reliable impact events to the cumulative damage is reduced, thus avoiding interference from incorrect attribution on the historical cumulative damage.

[0089] Step S53, calculate the damage amount of a single event. The historical cumulative damage of the impact-affected structural segment is superimposed. Obtain the updated cumulative damage. .

[0090] Step S6: Based on the updated cumulative damage, baseline deviation, and response change, obtain the fatigue level of the impact-affected structural segment and output the fatigue level. The fatigue level output method is not limited to displaying it on the host computer's screen, or sending it to a terminal or management platform.

[0091] Specifically, the weighted summation of the updated cumulative damage, baseline deviation, and response change yields a comprehensive risk index. : ; in, , , These are the user-configurable first fusion weight, second fusion weight, and third fusion weight, respectively. , , The range of values ​​is . Indicates baseline deviation. This represents the change in response. At this point, the following condition is met. .

[0092] Based on experience, multiple comprehensive risk index ranges corresponding to different fatigue levels are pre-set; comprehensive risk index The higher the value, the greater the degree of fatigue, and the higher the corresponding fatigue level. This is related to the comprehensive risk index. The fatigue level corresponding to the comprehensive risk index range into which the impact occurs is used as the fatigue level of the structural segment where the impact occurs.

[0093] In a preferred embodiment, before the structural section where the statistical impact occurs. The fatigue level of the structural segment where the impact occurred is obtained based on the percentage of peak response exceeding the limit during the second impact, the updated cumulative damage, baseline deviation, response change, and percentage of peak response exceeding the limit.

[0094] Specifically, before the structural segment where the statistical shock occurred. During a single impact, the number of times the peak response value in the structural segment where the impact occurred exceeds a preset peak response threshold. , , All are positive integers. The value range is from 5 to 20. The preset peak response threshold can be set according to the material of the impact-affected structural section; calculate the proportion of peak response exceeding the limit. , At this time, the comprehensive risk indicators : ; in, The fourth fusion weight is configurable by the user, and its value range is [value range missing]. At this point, satisfaction is achieved. This preferred embodiment addresses the issue of peak response exceeding the limit ratio. This allows structural sections that frequently experience peak response exceeding limits to obtain higher comprehensive risk indicators, thereby improving the accuracy of fatigue prediction.

[0095] Step S6 employs multiple dimensions to achieve a comprehensive quantitative assessment of structural fatigue risk, thereby improving the accuracy of fatigue early warning and its engineering applicability.

[0096] Example 2: The present invention also provides an embodiment of a processing device, including one or more processors and a memory; wherein the memory is used to store one or more computer programs, and the one or more processors are used to execute the one or more computer programs stored in the memory, so that the processors perform the features / steps of the above-described embodiment of the fatigue early warning method for logistics sorting production lines.

[0097] The embodiment is merely a specific example and does not indicate that this is the only way to implement the present invention.

[0098] The above description is merely a preferred embodiment of the present invention. Those skilled in the art will understand that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the present invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. A fatigue early warning method for logistics sorting production lines, characterized in that, The logistics sorting production line is divided into multiple structural sections, and nodes for structural monitoring are deployed on all or part of the structural sections. The method includes: The system receives event trigger messages uploaded by triggering nodes and auxiliary response features uploaded by one or more cooperating nodes. The event trigger message includes an impact time window and trigger response features. The structural segment where the cooperating node is located is connected to the structural segment where the triggering node is located. Both the trigger response features and the auxiliary response features include the impact arrival time, the response peak value, and the response spectrum distribution vector. The impact-generating structural segment and its actual response characteristics are determined based on the triggering response characteristics of the triggering node and the auxiliary response characteristics of the cooperating node. The source of the impact is determined based on the actual response characteristics of the impact-causing structural segment, the sorting task flow record, and the action flow record; the current working condition type is determined based on the sorting task flow record and the action flow record. The actual response characteristics of the impact-generating structural segment are compared with the baseline response characteristics and historical response characteristics of the impact-generating structural segment under the current working condition type and impact source type, respectively, to obtain the baseline deviation and response change. The cumulative damage amount of the impact-causing structural segment is updated based on the actual response characteristics. The fatigue level of the impact-causing structural segment is obtained based on the updated cumulative damage, the baseline deviation, and the response change, and the fatigue level is output. The determination of the impact-generating structural segment and its actual response characteristics based on the triggering response characteristics of the triggering node and the auxiliary response characteristics of the cooperating node includes: By comparing the trigger response characteristics of the triggering node with the auxiliary response characteristics of all the cooperating nodes, the actual arrival time difference sequence, the actual response peak attenuation ratio sequence, and the actual response spectrum similarity sequence between the triggering node and all the cooperating nodes are obtained. Multiple candidate structural segments are determined based on the structural topology diagram of the logistics sorting production line; For each candidate structural segment, the following steps are performed: Based on the simulated response database, obtain the theoretical arrival time difference sequence, theoretical response peak attenuation ratio sequence, and theoretical response spectrum similarity sequence of the triggering node and all the cooperating nodes when a simulated impact source is applied to the candidate structural segment; Based on the similarity between the actual arrival time difference sequence and the theoretical arrival time difference sequence, the similarity between the actual response peak attenuation ratio sequence and the theoretical response peak attenuation ratio sequence, and the similarity between the actual response spectrum similarity sequence and the theoretical response spectrum similarity sequence, obtain the localization score of the candidate structural segment; The impact-occurring structural segment is determined based on the location scores of all candidate structural segments; The actual response characteristics of the impact-causing structural segment are obtained based on the actual response characteristics of some nodes on the logistics sorting production line. The determination of the impact source type based on the actual response characteristics of the impact-generating structural segment, sorting task flow records, and action flow records includes: From the sorting task flow record and the action flow record, select one or more candidate impact sources that occur in the impact generation structure section. The candidate impact sources include at least one of the following: sorting tasks that transport packages through the impact generation structure section, grid actions in the impact generation structure section, and equipment start-up and shutdown actions. Calculate the time matching degree between each candidate impact source and the impact arrival time in the actual response characteristics; Calculate the peak response matching degree between each candidate impact source and the peak response in the actual response characteristics; The source confidence of each candidate impact source is obtained based on the time matching degree and the response peak matching degree; The candidate impact source with the highest source confidence is selected as the impact source, and the impact source type is determined.

2. The fatigue early warning method for logistics sorting production lines according to claim 1, characterized in that, The method further includes: statistical analysis of the structural segment where the impact occurs. The fatigue level of the impact-affected structural segment is obtained based on the percentage of peak response exceeding the limit during the second impact, the updated cumulative damage, the baseline deviation, the response change, and the percentage of peak response exceeding the limit. It is a positive integer.

3. The fatigue early warning method for logistics sorting production lines according to claim 1, characterized in that, Each node is equipped with an acceleration sensor and / or a strain sensor; when the acceleration detected by the acceleration sensor of a node is greater than an acceleration threshold, and / or when the strain detected by the strain sensor of a node is greater than a strain threshold, the node is designated as a trigger node, and the impact time window and trigger response features are extracted from the acceleration time series and / or strain time series.

4. The fatigue early warning method for logistics sorting production lines according to claim 1, characterized in that, The positioning score for each candidate structural segment is a weighted sum of the similarity between the actual arrival time difference sequence and the theoretical arrival time difference sequence, the similarity between the actual response peak attenuation ratio sequence and the theoretical response peak attenuation ratio sequence, and the similarity between the actual response spectrum similarity sequence and the theoretical response spectrum similarity sequence.

5. The fatigue early warning method for logistics sorting production lines according to claim 1, characterized in that, The step of determining the impact-causing structural segment based on the location scores of all candidate structural segments includes: sorting the candidate structural segments from high to low according to the location scores; if the difference in location scores between the first-ranked candidate structural segment and the second-ranked candidate structural segment is greater than a preset score difference threshold, then the first-ranked candidate structural segment is taken as the impact-causing structural segment.

6. The fatigue early warning method for logistics sorting production lines according to claim 1, characterized in that, Updating the cumulative damage of the impact-affected structural segment based on the actual response characteristics includes: Calculate the basic damage for a single event based on the actual response characteristics; The single-event basic damage is corrected using the source confidence of the impact source to obtain the single-event damage amount of the impact-occurring structural segment; The damage amount of the single event is added to the historical cumulative damage amount of the impact-occurring structural segment to obtain the updated cumulative damage amount.

7. The fatigue early warning method for logistics sorting production lines according to claim 1, characterized in that, The process of determining the current working condition type based on the sorting task flow record and the action flow record includes: Based on the sorting task flow record and action flow record, the logistics condition characteristics of the statistical period are extracted. The logistics condition characteristics include sorting task density, package weight distribution vector, proportion of heavy items, proportion of irregularly shaped items, conveying speed, frequency of compartment actions, and frequency of equipment start-up and shutdown actions. Calculate the similarity between the logistics condition features and the baseline logistics condition features of different condition types, and select the condition type with the highest similarity as the current condition type.

8. A processing apparatus, characterized in that, include: One or more processors; A memory for storing one or more computer programs, and one or more processors for executing the one or more computer programs stored in the memory to cause the one or more processors to perform the logistics sorting production line fatigue early warning method as described in any one of claims 1-7.

Images