Transmission equipment predictive maintenance algorithm

Abstract

The invention discloses a predictive maintenance algorithm for transmission equipment, and relates to the technical field of maintenance of the transmission equipment, in particular to a roller conveying system applied to flexible materials, which is characterized by comprising the following steps: S1, arranging position triggers at key transition nodes of a roller conveyor for acquiring timestamps of the materials passing through detection points in real time, generating a unique number and a corresponding time sequence relationship for each material; and S2, standard material passing time is calculated based on the theoretical beat period, the difference between the actual time and the standard period is calculated, and a beat offset set is constructed to reflect the conveying beat deviation of the materials between the adjacent detection points. According to the method, early warning and intelligent regulation and control of flexible material conveying rhythm imbalance are achieved through rhythm offset modeling, abnormal aggregation recognition and root cause diagnosis, compression bandwidth and driving time delay are dynamically adjusted when the imbalance risk occurs, the accumulation risk is effectively relieved, and the system stability, the automation level and the operation efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of transmission equipment maintenance, and particularly to a predictive maintenance algorithm for transmission equipment. Background Art

[0002] Predictive maintenance of transmission equipment refers to continuously monitoring data and performing intelligent analysis on automated material transmission equipment such as turnover machines and roller conveyors, to identify potential faults or performance degradation trends in advance, so as to achieve proactive intervention and planned maintenance before a fault occurs. This method usually involves the real-time collection of equipment operating state parameters (such as motor current, vibration spectrum, temperature rise curve, rotational speed fluctuation, etc.), and by constructing a health assessment model or introducing machine learning algorithms to judge the wear degree or failure probability of key components (such as bearings, chains, drums, etc.), thereby optimizing the maintenance timing, avoiding sudden shutdowns, improving transmission efficiency and production line stability, and reducing production interruptions and maintenance costs caused by faults.

[0003] The prior art has the following deficiencies: In the process of handling flexible materials such as soft packages or bags in the existing automated material conveying system, there is a technical problem of local accumulation imbalance due to insufficient control accuracy of the conveying rhythm. Especially at the corners, confluence sections or end buffer zones of the roller conveyor line, if there is a local imbalance in the roller drive rhythm, it is easy to cause the material aggregation density to rise rapidly within a short time, forming a "critical accumulation density" state. Since flexible materials themselves lack structural support capabilities, once the density critical point is exceeded, local collapse or mutual extrusion of materials will occur, resulting in content leakage and outer packaging damage, further inducing pushing imbalance and conveying blockage. Such problems not only damage the integrity of the materials, but may also cause chain-like accumulation blockages in severe cases, requiring manual intervention for disassembly and cleaning, which not only increases the risk of production line stoppage, but also significantly affects the stability and operation efficiency of the entire production line.

[0004] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] The object of the present invention is to provide a predictive maintenance algorithm for a transmission device. By collecting the material timestamps in real time at the key conveying nodes, constructing a beat offset model, combining density clustering and extreme value analysis to identify the out-of-tune aggregation segments, and using the distance change and rotational speed data for out-of-tune source diagnosis; calculating the beat out-of-tune index through the beat offset intensity, frequency and duration, and dynamically adjusting the compression bandwidth of the flexible material when the beat out-of-tune index exceeds the risk threshold, and realizing buffer adjustment by controlling the driving delay, effectively absorbing the beat disturbance, preventing material collapse and blockage, effectively improving the visualization management and adaptive ability of the conveying system, reducing the manual intervention and the risk of line stoppage, so as to solve the problems in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: A predictive maintenance algorithm for a transmission device, which is applied to a roller conveying system for flexible materials, and includes the following steps: S1. Arrange position triggers at each key transition node of the roller conveyor to collect the timestamps of the materials passing through the detection points in real time, and generate a unique number and the corresponding time series relationship for each material; S2. Calculate the standard material passing time based on the theoretical beat period, and calculate the difference between the actual time and the standard period to construct a set of beat offset amounts to reflect the conveying beat deviation of the materials between adjacent detection points; S3. Perform local extreme value analysis and density clustering operations on the set of beat offset amounts within a sliding time window, extract the out-of-tune aggregation sections, and mark their ranges and evolution trends on the time axis; S4. Based on the distance change curve of the front and rear materials, combined with the set rotational speed and feedback rotational speed data of each driving section, analyze whether the beat out-of-tune source is due to driving mismatch, material slip or foreign object interference; S5. Normalize the amplitude, frequency and duration of the beat offset, and combine the spatial position distribution weight of the out-of-tune section to calculate the beat out-of-tune index, which is used to quantify the synchronous stability level of the conveying system in the current cycle; S6. When the beat out-of-tune index exceeds the set risk threshold, dynamically adjust the compression bandwidth of the flexible material, and realize flexible stacking buffer and beat compensation by controlling the driving delay within the controllable stacking range.

[0007] Preferably, the step of arranging position triggers at each key transition node of the roller conveyor includes the following specific operations: At the key transition nodes of the roller conveyor, a position trigger is constructed by combining a laser pair - type sensor and a high - frequency encoder. This position trigger is used to accurately capture the moment when the front edge of the flexible material passes through with a high time resolution of a sampling interval less than 5 milliseconds, so as to achieve precise identification of the material passing at high speed. When the position trigger collects the timestamp signal of the material passing event, combined with the current operating state parameters of the roller conveyor section (including operating speed, drive number, etc.), a unique identifier is assigned to each unit of material. This identifier includes the material serial number, the section number where it is located, and timestamp information, ensuring the uniqueness and traceability of the material throughout the conveying process. The correspondence between the unique identifier and the timestamp is written into the central time - series cache table for subsequent beat - offset analysis and time - series modeling. This cache table has a first - in - first - out structure and supports real - time data update to ensure the timing consistency of the entire conveyor line and the ability to process high - frequency dynamic data.

[0008] Preferably, the step of calculating the standard material passing time based on the theoretical beat period and taking the difference between the actual time and the standard period to construct a set of beat offset amounts includes the following specific operations: Based on the structural parameters and rated drive speeds of each conveying section in the roller conveyor system, a theoretical beat - period model is established in advance. The model dynamically calculates the standard transmission time period that a unit of material should experience between the previous detection point and the current detection point under ideal conditions according to the length of each type of material, the length of the conveying section, and the set linear speed, and supports adaptive adjustment of model parameters for different flexible material characteristics (such as compressibility, slip tendency). Two timestamp values of the same - numbered material at adjacent detection points are retrieved from the central time - series cache table, and the difference between the two is used as the actual passing time of the material in this conveying section. This is then compared with the theoretical beat period of this section to calculate the beat offset value of the material, which is recorded in the set of beat offset amounts in the order of material arrival. To improve the sensitivity of identifying continuous beat disturbances, the system performs a sliding - section superposition algorithm on the set of beat offset amounts, applies weight marks to the part of the material flow with a continuous offset trend, and at the same time eliminates isolated abnormal offset values caused by single - point measurement errors or minor fluctuations in the conveying system, thereby constructing a set of beat offset amounts with trend stability and analysis robustness, providing a reliable data basis for subsequent clustering analysis, local synchronization disorder identification, and beat - disturbance modeling.

[0009] Preferably, the steps of performing local extreme value analysis and density clustering operation on the set of beat offset amounts within a sliding time window include the following operations: Set a sliding time window with a fixed length or dynamically adjustable. The system extracts a continuous data sequence from the set of beat offset amounts in a periodic scheduling manner as the local sample data within the current analysis window, which is used to capture the temporal fluctuation behavior of the beat in a short period. For the extracted local offset samples, perform local extreme value analysis operations. Adopt the first-order difference method combined with the sliding variance algorithm to identify the sudden rise, sudden drop, and critical oscillation points of the offset values, and extract the turning signals of beat disturbances and the inflection points of change trends from them to reveal potential starting points of synchronization disorders. Based on the density distribution characteristics of the offset values of this local sample, use a density-based spatial clustering algorithm (such as DBSCAN) to cluster and identify the beat offset points, automatically distinguish the aggregated segments of dense and continuous offsets from the outlier segments of scattered fluctuations, thereby constructing the precise boundaries and morphological characteristics of the beat disorder segments. Map the identified disorder-aggregated segments to the global time axis coordinate system, and perform evolution trend marking according to their duration, local density intensity, and offset trend slope within the sliding time window, classifying them into "growing", "decaying", or "fluctuating" disorder modes, providing classification judgment basis and dynamic input parameters for the system's synchronous speed regulation strategy selection and risk warning mechanism.

[0010] Preferably, map the identified disorder-aggregated segments to the global time axis coordinate system, and perform evolution trend marking according to the duration, local density intensity, and offset trend slope of the disorder-aggregated segments within the sliding time window, classifying them into "growing", "decaying", or "fluctuating" disorder modes. The specific steps are as follows: Map all the offset data points in the disorder-aggregated segments identified by density clustering back to the global time axis coordinate system according to their original timestamps, and construct the time range and start and end boundaries of this aggregated segment in the global dimension, so as to uniformly manage the disorder events within each local analysis window and achieve the time positioning and historical comparison of beat disturbances throughout the line. For the mapped disorder-aggregated segments, calculate their duration (the time span of the aggregated segment), local density intensity (the density or mean amplitude of offset points per unit time), and offset trend slope (the slope of the offset value change curve fitted based on the least squares method) within the sliding time window, comprehensively reflecting the intensity, stability, and development direction of this disorder segment. According to the preset multi-dimensional classification judgment rules, combine the above three quantitative indicators for trend classification: If the slope is positive and the density is increasing, it is marked as "growing"; if the slope is negative and the duration is lengthening, it is marked as "decaying"; if the slope fluctuates significantly but the overall amplitude is limited, it is classified as "fluctuating", and attach this evolution trend mark to the data element information of the disorder segment, providing strategy criteria and decision-making clues for subsequent segmented speed regulation strategies, priority control, and anomaly warning.

[0011] Preferably, the steps of analyzing the beat misalignment source based on the distance change curve of the front and rear materials and combining the set speed and feedback speed data analysis of each driving section include the following specific operations: Based on the timestamp information of the continuous material at the adjacent position trigger points, combined with the spatial structure parameters of the roller conveying section, dynamically calculate the material spacing change curve, and construct a "time-distance" relationship model through sliding fitting to detect whether there is a non-linear compression or expansion trend, revealing the dynamic evolution of the material flow tightness; Synchronously collect the set speed parameters and actual feedback speed signals of each driving section from the conveying control system, and compare them within the corresponding time window to identify whether there is a systematic deviation between the set and feedback, and further combine the material spacing trend to construct an association matrix between the driving response and the logistics behavior; Based on the constructed data model above, call the embedded multi-factor misalignment judgment logic, and judge one by one according to the three characteristic mode matching rules of driving mismatch, material slip, and foreign object interference: If the feedback speed of the driving section lags behind the set value abnormally, it is inferred as a driving mismatch; If the change in material spacing has no abnormal driving support, it is inferred as material slip; If there is a sudden spacing disturbance accompanied by a signal interruption, it is identified as foreign object interference, and the final diagnosis result is written into the misalignment event label for subsequent processing modules to call.

[0012] Preferably, the steps of calculating the beat misalignment index include the following operations: For each misalignment section identified within the sliding time window, sequentially extract the corresponding beat offset amplitude (i.e., the maximum offset value), offset frequency (the number of offsets per unit time), and duration (the span of the misalignment section on the time axis), and combine the system preset standard upper and lower limits to normalize the three-dimensional feature above to form a misalignment feature factor matrix under a unified scale, improving the horizontal comparability and modeling consistency of the misalignment states between multiple sections; Combine the physical space position of each misalignment section in the conveyor line, and perform spatial weight weighting on the normalized misalignment feature factors. For key control nodes such as corners, confluence sections, and end buffer areas, a higher sensitivity coefficient is assigned to reflect their importance in the overall stability structure of the system. Finally, form a comprehensive misalignment score that combines the beat disturbance intensity and the structural position weight; Use the comprehensive misalignment score as the input feature, import it into the beat misalignment index evaluation model, and adopt a weighted average mechanism or section aggregation rule to output a unified quantization index representing the beat synchronization stability within the current conveyor system cycle, that is, the beat misalignment index. This beat misalignment index can be used both as the core display parameter in the conveyor state visualization panel and as the criterion input for subsequent modules such as dynamic speed control, flexible buffer start-stop adjustment, and fault warning mechanism, realizing the accurate quantification and intelligent drive of the operating health status of the entire conveyor line.

[0013] Preferably, when the beat disorder index exceeds the set risk threshold, dynamically adjust the compression bandwidth of the flexible material and control the driving delay within the controllable stacking range. The specific steps are as follows: Continuously evaluate the beat disorder index of the conveyor line within each monitoring period and compare it with the preset risk threshold. When the beat disorder index is greater than the risk threshold, it indicates that the system beat disturbance has reached a perceptible risk level within the current period, and the flexible response mechanism needs to be activated to alleviate the disorder propagation. At this time, the system will dynamically calculate the compression bandwidth tolerance coefficient of the flexible material according to the current disorder intensity and distribution. The calculation formula is: , where: is the current beat disorder index, quantifying the synchronous offset degree of the roller conveyor system; is the set risk threshold, representing the lower limit of the allowable synchronous stability; is the historical maximum beat disorder index of the roller conveyor system, used for normalizing the adjustment scale; is the spatial sensitivity weight of the disorder section (such as giving a higher weight to the corner section); and are empirical adjustment coefficients, used to balance the influence of the global disorder intensity and structural risk; is the compression bandwidth tolerance coefficient, indicating the controllable stacking density range; After calculating the compression bandwidth tolerance coefficient, adjust the response delay time of the partition drive unit in the disorder aggregation section according to the compression bandwidth tolerance coefficient. The adjustment expression is: , where: is the dynamic delay response time of the drive unit; is the current average length of the flexible material, used to estimate the stacking space size; is the set conveying speed of the current section; is the response adjustment factor, used to control the response speed and stacking buffer sensitivity.

[0014] In the above technical solution, the technical effects and advantages provided by the present invention are: The present invention collects the time stamps of material movement in real time at key nodes of the conveying path, and constructs a beat offset model in combination with the theoretical beat, enabling the system to identify the trend of rhythm disorder at an early stage; further, density clustering and extreme value analysis are used to extract the disorder aggregation segments, and source classification diagnosis is carried out through the distance curve and rotational speed data to locate the root cause of the beat disturbance; finally, combined with the beat offset intensity, frequency and duration, a unified quantified "beat disorder index" is constructed, and when the risk threshold is exceeded, the controllable compression bandwidth of the flexible material is dynamically adjusted, and local buffering is achieved through the drive delay strategy, so as to effectively alleviate the disorder propagation and absorb the impact of beat fluctuations, and avoid problems such as collapse, extrusion, breakage or chain blockage of the flexible material in the key sections. This not only improves the visual management and self-adjustment ability of the conveying process, but also reduces the dependence on manual intervention, reduces the risk of system line stop, and has significant practical engineering application value and promotion prospects. Brief Description of the Drawings

[0015] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0016] Figure 1 It is a method flow chart of a predictive maintenance algorithm for a transmission device of the present invention. Detailed Embodiments

[0017] Now, the exemplary embodiments will be described more fully with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that the present disclosure will be more thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0018] The present invention provides a predictive maintenance algorithm for a transmission device as shown in Figure 1 and is applied to a roller conveying system for flexible materials, including the following steps: S1. Arrange position triggers at each key transition node of the roller conveyor to collect the time stamps of the material passing through the detection points in real time, and generate a unique number and corresponding time series relationship for each material; The step of arranging position triggers at each key transition node of the roller conveyor includes the following specific operations: At the key transition nodes of the roller conveyor, a position trigger is constructed by combining a laser pair sensor and a high-frequency encoder. This position trigger is used to accurately capture the moment when the front edge of the flexible material passes through with a high time resolution with a sampling interval of less than 5 milliseconds, so as to achieve precise identification of the material passing at high speed. After the position trigger collects the timestamp signal of the material passing event, combined with the current operating state parameters of the roller conveyor section (including operating speed, drive number, etc.), a unique identifier is assigned to each unit of material. This identifier includes the material serial number, the section number where it is located, and the timestamp information, ensuring the uniqueness and traceability of the material throughout the conveying process. The correspondence between the unique identifier and the timestamp is written into the central time series cache table for subsequent beat offset analysis and time series modeling. This cache table has a first-in-first-out structure and supports real-time data update to ensure the timing consistency of the entire conveyor line and the high-frequency dynamic data processing ability.

[0019] A position trigger is arranged at the key transition nodes of the roller conveyor and is used to collect the timestamp of the material passing through the detection point in real time, and generate a unique number and the corresponding time series relationship for each material. Its core role is to build a data foundation with high precision and high timing consistency for the entire set of predictive maintenance algorithms, so as to achieve precise modeling and tracking of the dynamic behavior of flexible materials during the automated conveying process. Since flexible packaging or bagged materials themselves lack structural stability, their motion states are extremely vulnerable to factors such as beat perturbation, uneven friction, and stacking resistance, resulting in the accumulation of small deviations in the running rhythm at certain nodes and evolving into serious faults such as stacking, collapse, or pushing imbalance. Therefore, it is necessary to perform fine-grained spatio-temporal marking on each unit of material to ensure that the system can master its actual passing time at each key node, so as to judge the beat consistency and transmission delay during the conveying process. By assigning a unique number to each material and constructing the corresponding relationship with the time series, not only the accurate identification and tracking of the material throughout the conveying path are realized, but also a reliable data foundation is provided for subsequent calculation of beat offset, identification of synchronization error, and analysis of drive misalignment.

[0020] S2. Calculate the standard material passing time based on the theoretical beat period, subtract the actual time from the standard period, and construct a set of beat offsets to reflect the conveying beat deviation of the material between adjacent detection points; The steps of calculating the standard material passing time based on the theoretical beat period, subtracting the actual time from the standard period, and constructing a set of beat offsets include the following specific operations: Based on the structural parameters and rated driving speeds of each conveying section in the roller conveying system, a theoretical beat cycle model is established in advance. The model dynamically calculates the standard transmission time cycle that a unit of material should experience between the previous detection point and the current detection point under ideal conditions according to the length of each type of material, the length of the conveying section, and the set linear speed, and supports the adaptive adjustment of model parameters for different flexible material characteristics (such as compressibility, slip tendency); two timestamp values of the same numbered material at adjacent detection points are retrieved from the central time series cache table, and the difference between the two is used as the actual passing time of the material in this conveying section. This is then compared with the theoretical beat cycle of this section, and the beat offset value of the material is calculated and recorded in the beat offset amount set in the order of material arrival; to improve the recognition sensitivity to continuous beat disturbances, the system performs a sliding section superposition algorithm on the beat offset amount set, applies weight marks to the part with a continuous offset trend in the material flow, and at the same time eliminates isolated abnormal offset values caused by single-point measurement errors or minor jitters of the conveying system, thereby constructing a beat offset amount set with trend stability and analysis robustness, providing a reliable data basis for subsequent clustering analysis, local synchronization disorder identification, and beat disturbance modeling.

[0021] The main function of this step is to reveal the small synchronization errors or potential maladjustment trends existing in the operation of the automated transmission system by quantifying the beat differences between the actual operation and the theoretical ideal state of the material during the conveying process, providing key basic data support for the predictive maintenance and adaptive control of the system. In the roller conveying system, especially in the scenario for flexible materials, due to the unstable physical form, easy slippage, and easy extrusion of flexible materials, the small disturbances they receive in the transmission path may quickly accumulate into problems such as beat disorders, local accumulations, or even overall blockages. Therefore, one cannot rely solely on the traditional "whether it arrives" logic, but must accurately calculate the conveying time of the material between each key detection node. By establishing a standard beat cycle model, the system can know how long each material should take to pass through adjacent detection points under ideal conditions, and taking the difference between this theoretical value and the actually measured time difference can form a set of "beat offset amounts", comprehensively reflecting the beat consistency and fluctuation characteristics between different conveying sections. This difference not only reveals the mismatch of local driving devices but may also reflect early signs of abnormal operations such as material accumulation, foreign object interference, and slip delay. As an important input for subsequent clustering analysis, beat disturbance trend identification, and speed regulation control strategies, the beat offset set can help the system achieve "continuous perception" and "dynamic correction" of the conveying state, thereby ensuring the timing coordination and flexible response ability of the whole line operation, and significantly reducing the failure rate and maintenance cost.

[0022] S3. Perform local extreme value analysis and density clustering operations on the set of beat offset amounts within the sliding time window, extract the out-of-tune aggregation section, and mark its range and evolution trend on the time axis; The steps of performing local extreme value analysis and density clustering operations on the set of beat offset amounts within the sliding time window include the following operations: Set a sliding time window with a fixed length or dynamically adjustable. The system extracts a continuous data sequence from the set of beat offset amounts in a periodic scheduling manner as the local sample data within the current analysis window to capture the temporal fluctuation behavior of the beat in a short period; For the extracted local offset samples, perform local extreme value analysis operations. Use the first-order difference method combined with the sliding variance algorithm to identify the sudden rise, sudden drop, and critical oscillation points of the offset value, and extract the turning signals and inflection points of the change trend of the beat disturbance from them to reveal potential starting points of synchronization out-of-tune; Based on the density distribution characteristics of the offset values of this local sample, use a density-based spatial clustering algorithm (such as DBSCAN) to cluster and identify the beat offset points, automatically distinguish the aggregated sections with dense and continuous offsets from the outlier sections with scattered fluctuations, so as to construct the precise boundaries and morphological characteristics of the beat out-of-tune section; Map the identified out-of-tune aggregation section to the global time axis coordinate system, and perform evolution trend marking according to its duration, local density intensity, and offset trend slope within the sliding time window, and classify it into "growing", "decaying", or "fluctuating" out-of-tune modes, providing classification judgment basis and dynamic input parameters for the system's synchronization speed regulation strategy selection and risk warning mechanism.

[0023] Map the identified dysregulation aggregation segments to the global time-axis coordinate system, and mark the evolution trends according to the duration, local density intensity, and offset trend slope of the dysregulation aggregation segments within the sliding time window, classifying them into "growing", "decaying", or "fluctuating" dysregulation patterns. The specific steps are as follows: Map all the offset data points in the dysregulation aggregation segments identified by density clustering back to the global time-axis coordinate system according to their original timestamps, construct the time range and start and end boundaries of the aggregation segment in the global dimension, so as to uniformly manage the dysregulation events within each local analysis window and achieve the time positioning and historical comparison of the beat disturbances throughout the line; For the mapped dysregulation aggregation segments, calculate their duration (aggregation segment time span), local density intensity (offset point density or mean amplitude per unit time), and offset trend slope (slope of the offset value change curve fitted based on the least squares method), comprehensively reflecting the intensity, stability, and development direction of the dysregulation segment; According to the preset multi-dimensional classification and determination rules, combine the above three quantitative indicators for trend classification: If the slope is positive and the density is increasing, mark it as "growing"; If the slope is negative and the duration is lengthening, mark it as "decaying"; If the slope fluctuates significantly but the overall amplitude is limited, classify it as "fluctuating", and attach the evolution trend mark to the data element information of the dysregulation segment to provide policy criteria and decision-making clues for subsequent segmented speed regulation strategies, priority control, and anomaly warnings.

[0024] Perform local extreme value analysis and density clustering operations on the set of beat offsets within a sliding time window, and extract the out-of-tune aggregation sections, marking their ranges and evolution trends on the time axis. The core function of this step is to achieve dynamic identification and trend perception of local beat synchronization disorders during the flexible material conveying process, thereby providing accurate basis for subsequent speed regulation strategies, risk early warnings, and maintenance decisions. Since in actual operation, beat offsets are often not linear or continuous abnormalities, but are manifested in the forms of "aggregated fluctuations", "local jitters", or "sudden drifts", with significant time locality and non-linear characteristics. If relying solely on global statistics or fixed threshold judgments, it is often impossible to detect these perturbation evolution processes with potential risks in a timely manner. Therefore, by means of a sliding time window, dynamic sampling analysis of beat offsets can be achieved. Combining local extreme value operations can detect the mutation points of beat perturbations, while through density clustering methods (such as DBSCAN), the boundaries and structures of abnormal aggregation paragraphs can be effectively identified, excluding the interference of noise and discrete points. Finally, by mapping the identified out-of-tune sections to the global time axis and classifying and labeling their evolution trends (such as growth, decay, fluctuations), the system can form the ability of time-series visualization and trend judgment of the beat synchronization state of the conveying chain, laying a data and model foundation for realizing flexible control and predictive maintenance. This step plays a connecting role in the entire predictive maintenance algorithm, is the key link from basic acquisition to intelligent identification, and is also the "perception center" for ensuring the stable conveying of flexible materials.

[0025] S4. Based on the distance change curve of the front and rear materials, combined with the set speed and feedback speed data of each drive section, analyze whether the beat offset source is due to drive mismatch, material slippage, or foreign object interference; Steps for analyzing the source of beat disorder based on the distance change curve of the front and rear materials and combining the set speed and feedback speed data of each drive section include the following specific operations: Based on the timestamp information of continuous materials at adjacent position trigger points, combined with the spatial structure parameters of the roller conveying section, dynamically calculate the material spacing change curve, and construct a "time-distance" relationship model through sliding fitting to detect whether there is a non-linear compression or expansion trend and reveal the dynamic evolution of the material flow tightness; Synchronously collect the set speed parameters and actual feedback speed signals of each drive section from the conveying control system, and compare them within the corresponding time window to identify whether there is a systematic deviation between the set and feedback, and further combine the material spacing trend to construct an association matrix between the drive response and the logistics behavior; Based on the constructed data model above, call the embedded multi-factor disorder judgment logic, and judge one by one according to the matching rules of three characteristic modes of drive mismatch, material slip, and foreign object interference: If the feedback speed of the drive section lags abnormally behind the set value, it is inferred as a drive mismatch; If the change in material spacing has no abnormal drive support, it is inferred as a material slip; If there is a sudden spacing disturbance accompanied by a signal interruption, it is identified as a foreign object interference, and the final diagnosis result is written into the disorder event label for subsequent processing modules to call.

[0026] Based on the distance change curve of the front and rear materials, combined with the set speed and feedback speed data of each drive section, analyzing whether the source of beat disorder is due to drive mismatch, material slip, or foreign object interference is the core step to achieve beat anomaly diagnosis and classification attribution. Its role is to achieve a leap from "phenomenon recognition" to "root cause positioning" by accurately identifying the causes of beat disorder, thereby providing a targeted and accurate intervention path for subsequent control strategies. In actual operation, due to its own unstable characteristics, flexible materials are extremely vulnerable to interference from minor abnormalities in the transmission system, such as drive motor response delay, belt slippage, sensor interference, or foreign object intrusion, and these factors often lead to the deviation of the conveying beat and trigger chain failures such as accumulation and collapse. Just relying on the beat deviation itself cannot determine which type of factor causes it. Therefore, it is necessary to introduce continuous modeling of the dynamic change of material spacing and conduct a comparative analysis in combination with the set value and actual feedback signal of each drive section during this period. By calculating the change trend of the front and rear material spacing, it can be identified whether there is a compression, drag, or abnormally sparse state in the logistics chain, and by correlating its analysis with the set-feedback deviation of the drive speed, it can further judge whether there are mode characteristics of "drive mismatch" (such as drive section lag or failure), "material slip" (such as the material's own displacement but normal drive), or "foreign object interference" (such as sudden jump or sensing interruption). The completion of this step enables the system to automatically classify and label different types of beat disturbance behaviors, which not only improves the dimension and accuracy of the system's perception ability but also provides a decision-making basis for hierarchical response strategies (such as speed regulation, obstacle clearing, alarm, etc.), and is a key support link for algorithm intelligence.

[0027] S5. Normalize the amplitude, frequency, and duration of the beat offset, and calculate the beat misalignment index in combination with the spatial position distribution weight of the misalignment section, which is used to quantify the synchronization stability level of the conveying system in the current cycle; The steps for calculating the beat misalignment index include the following operations: For each misalignment section identified within the sliding time window, sequentially extract the corresponding beat offset amplitude (i.e., the maximum offset value), offset frequency (the number of offsets per unit time), and duration (the span of the misalignment section on the time axis), and normalize the three-dimensional feature above in combination with the preset standard upper and lower limits of the system to form a misalignment feature factor matrix under a unified scale, improving the horizontal comparability and modeling consistency of the misalignment states among multiple sections; Combine the physical spatial position of each misalignment section in the conveyor line to perform spatial weight weighting on the normalized misalignment feature factors, where higher sensitivity coefficients are assigned to key control nodes such as corners, confluence sections, and end buffer areas to reflect their importance in the overall stability structure of the system, and finally form a comprehensive misalignment score that integrates the beat perturbation intensity and the structural position weight; Use the comprehensive misalignment score as an input feature and import it into the beat misalignment index evaluation model. Adopt a weighted average mechanism or a section aggregation rule to output a unified quantification index representing the beat synchronization stability in the current conveyor system cycle, that is, the beat misalignment index. This beat misalignment index can be used as both the core display parameter in the conveyor state visualization panel and the criterion input for subsequent modules such as dynamic speed control, flexible buffer start / stop adjustment, and fault warning mechanisms, achieving precise quantification and intelligent drive of the operating health status of the entire conveyor line.

[0028] Normalizing the amplitude, frequency, and duration of the beat offset and calculating the beat misalignment index in combination with the spatial position distribution weight of the misalignment section are key steps to realize the quantitative evaluation and intelligent decision support for the beat synchronization stability of the entire conveyor system in the current cycle. The core role of this step lies in breaking through the extensive method of the traditional conveyor system that can only judge the state by whether it is blocked or whether it times out, and instead introducing a multi-dimensional, continuous, and structure-aware index system to deeply model the beat perturbation and conduct trend evaluation, so as to achieve real-time health monitoring, predictive control, and adaptive optimization of adjustment strategies for the flexible material conveying system.

[0029] In the transmission of flexible packaging or bagged materials, a slight deviation in the beat is often an early signal before systematic failures (such as accumulation collapse, foreign object blockage, drive mismatch, etc.). A single indicator (such as a certain deviation time or a single lag) is difficult to reflect the overall state of the system. Therefore, in this step, three core features of the misalignment section are first extracted: deviation amplitude (fluctuation intensity), deviation frequency (density of disturbances), and duration (trend of abnormal continuation), and normalization processing is performed, so that the deviation behaviors under different time windows and different material types can be compared and modeled under the same dimension. This step not only standardizes the data but also improves the stability and generalization ability of subsequent analysis and modeling.

[0030] Furthermore, to overcome the evaluation misunderstanding that "all misalignment events have the same weight", this step introduces a spatial position distribution weight mechanism to consider the influence degree of different conveying sections on the system stability. For example, beat misalignment occurring at system bottleneck positions such as corner sections, confluence points, and buffer areas will bring more serious chain consequences. Therefore, higher weights should be assigned to the disturbance factors corresponding to these key nodes. This spatial sensitivity mechanism makes the beat misalignment index not only reflect "how serious the beat disturbance is" but also embody "how important this disturbance is to the whole line", improving the structure perception ability of the algorithm.

[0031] Finally, through the weighted integration operation of the normalization factor and spatial weight of each misalignment section, a unified "beat misalignment index" is output as the core index reflecting the synchronous stability of the current cycle's conveying. This beat misalignment index can be used to drive the speed control logic of the production management system, flexible stacking control strategies, warning system criteria, etc., to achieve visual quantification and intelligent feedback on the operating state of the conveying system, provide a clear stability rating basis for on-site engineers, and can also be used as one of the input parameters of the predictive maintenance system to predict the evolution trend of potential risks, ultimately improving the operating safety, flexible response ability, and intelligent control level of the entire conveying line.

[0032] S6. When the beat misalignment index exceeds the set risk threshold, dynamically adjust the compression bandwidth of the flexible material, and realize flexible stacking buffering and beat compensation by controlling the drive delay within the controllable stacking range; When the beat misalignment index exceeds the set risk threshold, dynamically adjust the compression bandwidth of the flexible material and control the drive delay within the controllable stacking range. The specific steps are as follows: Continuously evaluate the beat misalignment index of the conveying line in each monitoring cycle and compare it with the preset risk threshold. When it is satisfied that the beat misalignment index is greater than the risk threshold, it means that the system beat disturbance has reached the perceivable risk level in the current cycle, and the flexible response mechanism needs to be activated to relieve the misalignment propagation. At this time, the system will dynamically calculate the compression bandwidth tolerance coefficient of the flexible material according to the current misalignment intensity and distribution. The calculation formula is: , where: is the current beat imbalance index, quantifying the degree of synchronous offset of the roller conveyor system; is the set risk threshold, representing the lower limit of allowable synchronous stability; is the historical maximum beat imbalance index of the roller conveyor system, used for normalizing the adjustment scale; is the spatial sensitivity weight of the imbalance section (e.g., higher weight is given to the corner section); and are empirical adjustment coefficients, used to balance the influence of global imbalance intensity and structural risk; is the compression belt tolerance coefficient, representing the controllable stacking density range; The function of this step is to obtain a regulation factor for the controllable flexible stacking compression range by quantifying the beat imbalance intensity and linking the spatial structure sensitivity, which serves as the input basis for the subsequent drive regulation strategy.

[0033] After calculating the compression belt tolerance coefficient, the response delay time of the partition drive unit is adjusted within the imbalance aggregation section according to the compression belt tolerance coefficient. The adjustment expression is: , where: is the dynamic delay response time of the drive unit; is the current average length of the flexible material, used to estimate the stacking space size; is the set conveying speed of the current section; is the response adjustment factor, used to control the response speed and stacking buffer sensitivity; This delay will be distributed to multiple drive controllers, causing instantaneous beat "lag" behavior in some drive sections, artificially creating controllable material stacking, and using the flexibility and compressibility of the material to eliminate beat offset fluctuations, realizing the "dynamic lag - buffer absorption" mechanism. The function of this step is to convert the calculated flexible response strategy into specific drive instructions, achieve real-time dynamic compensation for the beat synchronization state, and at the same time avoid forming uncontrollable congestion or material damage, ensuring the stability and flexible robustness of the whole line operation.

[0034] Adjusting the compression bandwidth of flexible materials is essentially achieved by controlling the controllable stacking degree of materials in a local area of the conveyor line to buffer the material flow fluctuations caused by beat disorders. This goal can be achieved in various ways: First, the rotational speed or start-stop delay of the driving rollers can be dynamically adjusted to make the materials stay in a certain section for a slightly longer time, so as to achieve flexible stacking without actual blockage. Second, the inter-segment control logic can be adjusted according to the real-time beat state. For example, the start of the downstream segment can be delayed and the upstream segment can be paused in advance to create a space compression area in terms of rhythm. Third, the buffer or transition area designed in the conveying path can be utilized to stack the materials orderly in this area, ensuring that the stacking density is controlled within the structural limit that the packaging materials can bear. In addition, the intelligent recognition of the material flow state and the calculation of the compression bandwidth can be carried out at the software level by synchronously controlling the sensor trigger frequency and the material number tracking algorithm.

[0035] By integrating these methods, without physical modification of the system, a dynamic stacking bandwidth adjustment mechanism of "safe, predictable, and controllable" for flexible materials in the conveying system can be achieved through the combined regulation of software and hardware, so as to effectively absorb beat disturbances and improve the system stability and intelligent response ability.

[0036] When the beat disorder index exceeds the set risk threshold, the compression bandwidth of flexible materials is dynamically adjusted, and flexible stacking buffering and beat compensation are achieved by controlling the driving delay within the controllable stacking range. Its core function is to provide an adaptive buffering and adjustment mechanism for the conveying system to offset the material flow fluctuations caused by beat synchronization disorders, thereby avoiding serious consequences such as stacking collapse, jamming, and chain blockage, and ensuring the continuity and stability of the entire conveyor line. The logic behind this step is based on an in-depth understanding of the physical properties and system behavior laws in the process of flexible material conveying: Flexible materials such as bagged and soft-packaged products have unstable structures and are prone to over-dense stacking and even damage due to local rhythm abnormalities, while traditional conveying systems often lack fine-grained buffering and adjustment capabilities. Once beat disorders occur, they are likely to quickly evolve into large-scale failures.

[0037] By calculating the beat disorder index and judging whether it exceeds the risk threshold, the system can trigger the response mechanism at the initial stage of identifying the disorder trend. The so-called "compression bandwidth" refers to the behavior of allowing a certain degree of "controllable stacking" of flexible materials in a short time and limited space in the conveying path. This kind of stacking is different from congestion and is a "soft buffer layer" under system perception and algorithm regulation. The size of the compression bandwidth is dynamically calculated through the compression bandwidth tolerance coefficient, which reflects the tolerance range of the system for the current beat risk. This parameter will affect the response logic of each downstream driving segment. If the current beat disorder index is high and the compression bandwidth tolerance coefficient increases, the system will allow the driver to start with a delay and create a "delay zone" in a certain section, prompting the flexible materials to automatically stack in this section, thereby absorbing the impact brought by the upstream beat fluctuations.

[0038] The setting of the driving delay not only considers the length, conveying speed, and density distribution of the current material, but also refers to the location and evolution trend of the imbalance occurrence, enabling the adjustment to have spatial perception ability and dynamic response ability. In this way, when the system detects abnormal rhythm in a certain area, it no longer passively relies on sensor alarms or line stops, but actively adjusts the driving rhythm. Through the "buffer absorption + synchronous compensation" mechanism, it locally alleviates the impact of the imbalance while maintaining the continuity and stability of the material flow of the entire line.

[0039] Generally speaking, the role of this step is to construct an intelligent and adaptive stacking buffer strategy in the flexible material conveying environment, replacing the static transmission logic with a dynamic control strategy, enabling the system to "flexibly respond, smoothly transition, and automatically repair" when facing beat disturbances, greatly improving the fault tolerance, robustness, and operating efficiency of the flexible conveying system, and providing a core control means for realizing intelligent predictive maintenance.

[0040] Through the implementation of the above-mentioned predictive maintenance algorithm for the conveying equipment, it is possible to achieve fine perception, dynamic recognition, and adaptive regulation of the beat synchronization state during the flexible material conveying process, significantly enhancing the operating stability and fault prevention ability of the entire roller conveying system. This solution constructs a beat offset model by collecting the material movement timestamps at the key nodes of the conveying path in real time and combining with the theoretical beat, enabling the system to identify the trend of rhythm imbalance in the early stage; further, by means of density clustering and extreme value analysis, the imbalance aggregation section is extracted, and the root cause classification diagnosis is carried out through the distance curve and rotational speed data to achieve the root cause location of the beat disturbance; finally, combined with the beat offset intensity, frequency, and duration, a unified quantified "beat imbalance index" is constructed, and when it exceeds the risk threshold, the controllable compression bandwidth of the flexible material is dynamically adjusted, and local buffering is achieved through the driving delay strategy, thereby effectively alleviating the spread of the imbalance and absorbing the impact of beat fluctuations, and avoiding problems such as collapse, extrusion, damage, or chain blockage of the flexible material in the key sections. In summary, this algorithm not only improves the visual management and self-regulation ability of the conveying process, but also reduces the dependence on manual intervention, reduces the risk of system line stops, and has significant practical engineering application value and promotion prospects.

[0041] The above formulas are all dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0042] Only some exemplary embodiments of the present invention have been described by way of illustration. Without doubt, for those of ordinary skill in the art, various different ways can be used to modify the described embodiments without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0043] As described above, the foregoing is only a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be subject to the scope of protection of the claims.

Claims

1. A predictive maintenance algorithm for a transmission device, which is applied to a roller conveyor system for flexible materials, and is characterized in that, The steps include: S1. Position triggers are arranged at each key transition node of the roller conveyor to collect the timestamp of the material passing through the detection point in real time, and generate a unique number and corresponding time series relationship for each material; S2. Calculate the standard material passing time based on the theoretical beat cycle, and subtract the actual time from the standard cycle to construct a beat offset set to reflect the conveying beat deviation of the material between adjacent detection points; S3, performing local extreme value analysis and density clustering operations on the beat offset set within the sliding time window, extracting the disorder clustering segment, and marking its range and evolution trend on the time axis; S4. Based on the distance change curve of the front and rear materials, combined with the set speed and feedback speed data of each drive section, analyze whether the beat imbalance source is caused by drive mismatch, material slippage or foreign body interference; S5, normalizing the amplitude, frequency and duration of the beat deviation, and calculating the beat deviation index in combination with the spatial position distribution weight of the deviation section, so as to quantify the synchronization stability level of the conveying system in the current cycle; S6. When the beat disorder index exceeds the set risk threshold, the compression bandwidth of the flexible material is dynamically adjusted, and flexible stacking buffering and beat compensation are achieved by controlling the driving delay within the controllable stacking range.

2. The predictive maintenance algorithm for a transmission device according to claim 1, wherein The steps for placing position triggers at each key transition node of the roller conveyor include the following specific operations: A position trigger is constructed by combining a laser beam sensor and a high-frequency encoder at the key transition node of the roller conveyor. The position trigger is used to capture the moment when the front edge of the flexible material passes with a high time resolution of less than 5 milliseconds at a sampling interval. When the position trigger collects the timestamp signal of the material passing event, it assigns a set of unique identifiers to each unit of material in combination with the current operating status parameters of the roller conveyor section. The identifier includes the material serial number, the section number and the timestamp information. The correspondence between the unique identifier and the timestamp is written into the central time series cache table for subsequent beat offset analysis and time series modeling.

3. The predictive maintenance algorithm for a transmission device according to claim 2, characterized in that, Calculate the standard material passing time based on the theoretical beat cycle, and subtract the actual time from the standard cycle to construct the beat offset set as follows: Based on the structural parameters and rated drive speed of each conveying section in the roller conveyor system, a theoretical beat cycle model is pre-established. According to the length of each type of material, the length of the conveying section and the set line speed, the standard transmission time cycle that a unit material should experience from the previous detection point to the current detection point under ideal conditions is dynamically calculated; Retrieve two timestamp values of the same numbered materials at adjacent detection points from the central time series cache table, take the difference between the two as the actual passing time of the material in the conveying section, and compare it with the theoretical beat cycle of the section to calculate the beat offset value of the material, and record it in the beat offset set according to the order of material arrival; The sliding segment superposition algorithm is performed on the beat offset set, and weight marks are applied to the parts of the material flow with continuous offset trends. At the same time, isolated abnormal offset values caused by single-point measurement errors or slight jitters of the conveying system are eliminated.

4. The predictive maintenance algorithm for a transmission device according to claim 1, wherein The steps of performing local extreme value analysis and density clustering operation on the set of beat offset amounts within a sliding time window are as follows: Set a sliding time window with a fixed length or dynamically adjustable, and extract a continuous data sequence from the set of beat offset amounts in a periodic scheduling manner as the local sample data within the current analysis window, which is used to capture the temporal fluctuation behavior of the beat in a short period; For the extracted local offset samples, perform local extreme value analysis operations. Adopt the first-order difference method combined with the sliding variance algorithm to identify the sudden increase, sudden decrease, and critical oscillation points of the offset values, and extract the turning signals of beat disturbances and the inflection points of change trends therefrom; Based on the density distribution characteristics of the offset values of this local sample, use the density-based spatial clustering algorithm to cluster and identify the beat offset points, automatically distinguish the aggregated segments with dense continuous offsets from the outlier segments with scattered fluctuations, and construct the precise boundaries and morphological characteristics of the beat out-of-tune sections; Map the identified out-of-tune aggregated sections to the global time axis coordinate system, and perform evolution trend marking according to their duration, local density intensity, and offset trend slope within the sliding time window, and classify them into "growing", "decaying", or "fluctuating" out-of-tune modes.

5. The predictive maintenance algorithm for a transmission device according to claim 4, wherein Map the identified out-of-tune aggregated sections to the global time axis coordinate system, and perform evolution trend marking according to the duration, local density intensity, and offset trend slope of the out-of-tune aggregated sections within the sliding time window, and classify them into "growing", "decaying", or "fluctuating" out-of-tune modes. The specific steps are as follows: Map all the offset data points in the out-of-tune aggregated sections identified by density clustering back to the global time axis coordinate system according to their original timestamps, and construct the time range and start and end boundaries of this aggregated section in the global dimension; For the mapped out-of-tune aggregated sections, calculate their duration, local density intensity, and offset trend slope within the sliding time window; According to the preset multi-dimensional classification and determination rules, combine the quantization indicators for trend classification: if the slope is positive and the density is increasing, mark it as "growing"; if the slope is negative and the duration is lengthening, mark it as "decaying"; if the slope fluctuates significantly but the overall amplitude is limited, classify it as "fluctuating", and attach this evolution trend mark to the data element information of the out-of-tune section.

6. The predictive maintenance algorithm for a transmission device according to claim 1, wherein The steps of analyzing the beat out-of-tune source based on the distance change curve of the front and rear materials and combining the set speed and feedback speed data of each driving section are as follows: Based on the timestamp information at the adjacent position trigger points of the continuous materials, combine the spatial structure parameters of the roller conveying section, dynamically calculate the material spacing change curve, and construct a "time-distance" relationship model through sliding fitting; Synchronously collect the set speed parameters and actual feedback speed signals of each driving section from the conveying control system, and compare them within the corresponding time window to identify whether there is a systematic deviation between the set and feedback, and further combine the material spacing trend to construct an association matrix between the driving response and the logistics behavior; Based on the constructed data model, call the embedded multi-factor imbalance judgment logic and judge one by one according to the matching rules of three characteristic patterns: drive mismatch, material slip, and foreign object interference. If the feedback speed of the drive section lags abnormally behind the set value, it is inferred as a drive mismatch. If the change in material spacing has no abnormal drive support, it is inferred as material slip. If there is a sudden spacing disturbance accompanied by signal interruption, it is identified as foreign object interference.

7. The predictive maintenance algorithm for a transmission device according to claim 1, wherein The steps to calculate the beat imbalance index are as follows: For each imbalance section identified within the sliding time window, sequentially extract the corresponding beat offset amplitude, offset frequency, and duration for each section, and combine the preset upper and lower limits of the roller conveyor system standard to normalize the three-dimensional features, forming an imbalance feature factor matrix under a unified scale. Combined with the physical space position of each imbalance section in the conveyor line, perform spatial weight weighting on the normalized imbalance feature factors, where a higher sensitivity coefficient is assigned to key control nodes, forming a comprehensive imbalance score that combines the beat disturbance intensity and structural position weight. Use the comprehensive imbalance score as an input feature and import it into the beat imbalance index evaluation model. Adopt a weighted average mechanism to output a unified quantization index representing the beat synchronization stability within the current conveyor system cycle, that is, the beat imbalance index.

8. The predictive maintenance algorithm for a transmission device according to claim 1, wherein When the beat imbalance index exceeds the set risk threshold, dynamically adjust the compression bandwidth of the flexible material and control the drive delay within the controllable accumulation range. The specific steps are as follows: Continuously evaluate the beat imbalance index of the conveyor line within each monitoring period, and compare it with the preset risk threshold for judgment. When the beat imbalance index is greater than the risk threshold, activate the flexible response mechanism to alleviate the spread of imbalance. Dynamically calculate the compression zone tolerance coefficient of the flexible material according to the current imbalance intensity and distribution. The calculation formula is: , where: is the current beat imbalance index, quantifying the synchronous offset degree of the roller conveyor system; is the set risk threshold, representing the lower limit of allowable synchronous stability; is the historical maximum beat imbalance index of the roller conveyor system, used for normalizing the adjustment scale; is the spatial sensitivity weight of the imbalance section; and are empirical adjustment coefficients, used to balance the influence of global imbalance intensity and structural risk; is the compression zone tolerance coefficient, indicating the controllable stacking density range; After calculating the compression belt tolerance coefficient, the response delay time of the partition drive unit is adjusted according to the compression belt tolerance coefficient within the misalignment aggregation section. The adjustment expression is as follows: , where: is the dynamic delay response time of the drive unit; is the average length of the current flexible material, which is used to estimate the size of the stacking space; is the set conveying speed of the current section; is the response adjustment factor, which is used to control the response speed and the stacking buffer sensitivity.

Images