Self-diagnosis and early warning method for operation state of carton conveying equipment

Abstract

The invention discloses a self-diagnosis and early warning method for the running state of carton conveying equipment, and belongs to the technical field of industrial automation control. The method comprises the steps of collecting a vibration spectrum, a current ripple fluctuation ratio, equipment operation parameters, a vibration wave conduction time sequence and laser scanning; generating vibration current composite sensing data and carton deformation characteristics; establishing an equipment dynamic health file; carrying out abnormal separation on the composite sensing data and the health file to obtain a primary diagnosis result; based on the result, fusing a vibration wave time sequence and deformation characteristics to execute fault tracing, and outputting secondary positioning information; and generating a maintenance decision instruction according to the positioning information, and outputting an early warning containing a fault component identifier and a maintenance measure. According to the method, the multi-source data fusion and intelligent diagnosis algorithm is adopted, and the dynamic health archives are combined, so that the equipment abnormity can be accurately positioned, and the equipment reliability and the maintenance efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial automation control, and in particular to a self-diagnosis and early warning method for the operating status of carton conveying equipment. Background Art

[0002] Carton conveying equipment is an essential component of automated production lines and logistics warehousing systems. Its primary function is to ensure the continuous, efficient, and stable transportation of cartons and other materials. In the context of modern industry and intelligent manufacturing, this type of equipment is often integrated into complex flexible manufacturing systems. The reliability of its operating status directly impacts the efficiency and cost of the entire production or logistics process. Therefore, real-time operating status monitoring and fault warning of carton conveying equipment are crucial technical support for ensuring smooth production line operation and implementing predictive maintenance.

[0003] Existing methods for monitoring the condition of carton conveying equipment often rely on the detection of a single physical quantity. For example, these methods involve installing vibration sensors to monitor the mechanical vibration of key components such as bearings and motors, or monitoring the current of drive motors to determine load anomalies. While some systems collect multiple signals simultaneously, these signals are typically analyzed independently at the data processing level, failing to achieve deep information fusion. When a metric is detected exceeding a preset threshold, the system issues an alarm, requiring maintenance personnel to rely on their personal experience to determine the specific cause and location of the fault.

[0004] The above-mentioned existing technologies have obvious deficiencies in practical applications. First, the ability of single signal monitoring to identify complex faults is limited. The early signs of many faults are the result of the coupling of multiple physical quantities, which are difficult to be effectively captured by a single signal. Secondly, independent signal analysis methods ignore the inherent correlation between different physical phenomena, which can easily lead to false alarms or missed alarms, and are not very accurate in locating the source of the fault, and can often only provide a vague fault area. In addition, traditional alarm methods can only inform equipment of abnormalities, and cannot provide specific fault diagnosis conclusions and maintenance guidance. They are extremely dependent on the professional skills and experience of maintenance personnel, resulting in low troubleshooting efficiency and increased unplanned downtime. Summary of the Invention

[0005] To solve the above problems, the present invention provides a self-diagnosis and early warning method for the operating status of carton conveying equipment. It adopts multi-source data fusion and intelligent diagnosis algorithm, combined with dynamic health records, to achieve accurate positioning of equipment abnormalities and improve equipment reliability and maintenance efficiency.

[0006] The above objectives can be achieved through the following solutions: A self-diagnosis and early warning method for the operating status of carton conveying equipment includes collecting vibration spectrum, current ripple fluctuation rate, equipment operating parameters, vibration wave conduction time sequence, and laser scanning; generating vibration current composite sensor data and carton deformation characteristics; establishing a dynamic health file for the equipment; performing anomaly separation on the composite sensor data and health file to obtain a primary diagnosis result; based on the result, the vibration wave time sequence and deformation characteristics are integrated to perform fault tracing and output secondary positioning information; generating maintenance decision instructions based on the positioning information, and outputting a warning containing the fault component identification and maintenance measures.

[0007] Optionally, generating composite sensing data based on the vibration spectrum data and the current ripple fluctuation rate data includes: performing time domain alignment processing on the vibration spectrum data and the current ripple fluctuation rate data to generate a time domain synchronization feature vector; extracting the energy gradient change characteristics in the time domain synchronization feature vector to generate composite sensing data.

[0008] Optionally, the time domain alignment processing of the vibration spectrum data and the current ripple fluctuation rate data to generate a time domain synchronization feature vector includes: obtaining a time domain vibration sequence based on the vibration spectrum data; performing time domain waveform normalization processing on the current ripple fluctuation rate data to obtain a current fluctuation time series; and performing time domain alignment processing on the time domain vibration sequence and the current fluctuation time series to generate a time domain synchronization feature vector.

[0009] Optionally, the fault tracing analysis includes: obtaining the conduction threshold and vibration peak timing data of the vibration sensor; calculating the first difference in propagation delay based on the vibration peak timing data; when the first difference in propagation delay exceeds the conduction threshold, the positioning motor bearing fails.

[0010] Optionally, generating carton deformation features based on the laser scanning data includes: performing dynamic displacement compensation and geometric contour extraction on the laser scanning data to obtain a calibrated three-dimensional edge coordinate set; performing deformation physical quantity decoupling calculation on the three-dimensional edge coordinate set to generate carton deformation features.

[0011] Optionally, establishing a dynamic health profile of the equipment includes: obtaining current harmonic amplitude ratio data from the equipment operating parameters; monitoring the current harmonic amplitude ratio data to obtain a time-varying degradation offset; and establishing a dynamic health profile of the equipment based on the time-varying degradation offset.

[0012] Optionally, generating a maintenance decision instruction based on the secondary positioning information includes: performing fault topology modeling on the secondary positioning information to obtain a fault positioning topology map; performing maintenance strategy matching and optimization on the fault positioning topology map to generate a maintenance decision instruction.

[0013] Optionally, the output of the warning signal containing the fault component identification and maintenance measures based on the maintenance decision instruction includes: parsing the maintenance decision instruction to obtain the optimization strategy parameters of the fault component maintenance measures; performing adaptive signal encapsulation processing on the optimization strategy parameters of the fault component maintenance measures, and outputting the warning signal containing the fault component identification and maintenance measures.

[0014] Optionally, the maintenance strategy matching and optimization of the fault location topology map to generate maintenance decision instructions includes: performing node correlation analysis and maintenance path weight calculation on the fault location topology map to obtain a path weight matrix; performing dynamic strategy matching and maintainability index optimization based on the path weight matrix to generate maintenance decision instructions.

[0015] Based on the same inventive concept, the present invention also provides a self-diagnosis and early warning system for the operating status of a carton conveying equipment, the system comprising: a multi-source signal acquisition module for acquiring vibration spectrum data, current ripple fluctuation rate data, equipment operating parameters, vibration wave conduction timing data and laser scanning data; a composite sensing generation module for generating composite sensing data based on the vibration spectrum data and current ripple fluctuation rate data; a deformation feature extraction module for generating carton deformation features based on the laser scanning data; a health record modeling module for establishing a dynamic health record of the equipment based on the equipment operating parameters; an abnormality diagnosis module for performing abnormality separation processing on the composite sensing data and the dynamic health record of the equipment to generate a primary diagnosis result; a fault tracing module for performing fault tracing analysis on the vibration wave conduction timing data and the carton deformation features based on the primary diagnosis result to generate secondary positioning information; a maintenance decision module for generating a maintenance decision instruction according to the secondary positioning information; and an early warning output module for outputting an early warning signal containing a fault component identification and maintenance measures based on the maintenance decision instruction.

[0016] Compared with the prior art, the present invention has the following advantages: This invention generates composite sensor data by fusing vibration spectrum data with current ripple fluctuation data, enabling coordinated monitoring of equipment status from both mechanical and electrical perspectives. This multi-physics fusion approach can capture early signs of faults that are difficult to detect with a single signal, such as mechanical vibrations caused by electrical system anomalies. This improves the sensitivity and foresight of fault diagnosis, enabling effective early warning of faults.

[0017] This method innovatively incorporates vibration wave transmission time series data and carton deformation characteristics for fault tracing and analysis. By measuring the propagation delay of vibration signals between key equipment components and combining this with laser scanning to analyze the deformation pattern of the conveyed material itself, it achieves dual verification and precise location of the fault source. This method can clearly distinguish between faults in different components, such as motor bearing failure and loose tensioner bolts, improving the accuracy and reliability of fault diagnosis.

[0018] This invention establishes a complete intelligent process from fault location to maintenance decision-making. After accurately locating the fault, the system automatically models the fault topology and optimizes the maintenance strategy, generating maintenance decision instructions that identify the specific faulty component and optimize the maintenance measures. This transforms early warning information from a simple alarm signal into a directly executable intelligent maintenance plan, improving maintenance efficiency and reducing equipment downtime.

[0019] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 The present invention is a flowchart of a method for self-diagnosis and early warning of the operating status of carton conveying equipment according to an embodiment of the present invention.

[0022] Figure 2 4 is a diagram of the timing data of vibration wave conduction according to an embodiment of the present invention.

[0023] Figure 3 3 is a comparison diagram of the theoretical deformation of cartons under load according to an embodiment of the present invention.

[0024] Figure 4 This is a dynamic health profile diagram of a device according to an embodiment of the present invention.

[0025] Figure 5 This is a device dynamic health profile degradation curve diagram according to an embodiment of the present invention.

[0026] Figure 6 The present invention is a schematic structural diagram of a carton conveying equipment operating status self-diagnosis and early warning system. DETAILED DESCRIPTION

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0028] Reference Figure 1 One embodiment of the present invention proposes a self-diagnosis and early warning method for the operating status of carton conveying equipment. By adopting multi-source data fusion and intelligent diagnosis algorithm, combined with dynamic health records, it can accurately locate equipment abnormalities and improve equipment reliability and maintenance efficiency.

[0029] The method of this embodiment specifically includes: Obtain vibration spectrum data, current ripple fluctuation rate data, equipment operating parameters, vibration wave transmission timing data and laser scanning data; generating composite sensing data based on the vibration spectrum data and the current ripple fluctuation rate data; generating a carton deformation feature based on the laser scanning data; Establishing a dynamic health profile of the equipment based on the equipment operating parameters; Performing abnormality separation processing on the composite sensor data and the dynamic health record of the device to generate a primary diagnosis result; Based on the first-level diagnostic results, a fault tracing analysis is performed on the vibration wave transmission time series data and the carton deformation characteristics to generate second-level positioning information; generating maintenance decision instructions based on the secondary positioning information; Based on the maintenance decision instruction, an early warning signal including a fault component identification and maintenance measures is output.

[0030] Specifically, firstly, the mechanical vibration, driving current, macroscopic operation parameters, vibration wave transmission time series and three-dimensional morphological data of the transported object are synchronously collected. The vibration wave transmission time series data is as follows: Figure 2As shown, all-round perception of the equipment status is achieved. The core is to deeply integrate the vibration spectrum and current ripple that reflect the subtle dynamics inside the equipment to generate a composite sensor data that is more sensitive to early faults. At the same time, by continuously monitoring the equipment operating parameters, a dynamically changing health baseline is established, that is, the equipment dynamic health file, which serves as the basis for distinguishing normal aging from abnormal degradation. The diagnostic process adopts a hierarchical strategy. The first-level diagnosis uses composite sensor data to compare with the health file to quickly determine whether there is an abnormality. It is achieved through dynamic statistical thresholds. The latest 30 time-varying degradation offset samples are extracted from the equipment dynamic health file every second, and their average is calculated. and standard deviation If the current energy gradient value of the composite sensor data exceeds , it is determined to be a Level 1 anomaly. Once the anomaly is confirmed, Level 2 diagnosis—fault source analysis—is initiated. This analysis innovatively combines the time delay characteristics of vibration waves propagating through the equipment structure with the physical deformation characteristics of the cardboard box acquired through laser scanning to achieve precise spatial localization of the fault source. Finally, based on this precise positioning information, the system intelligently generates maintenance decision instructions containing specific component identification and optimized repair steps through fault topology modeling and maintenance strategy optimization. This is then packaged as an early warning signal output, forming a complete technical closed loop from state perception, anomaly detection, fault location, to maintenance decision-making.

[0031] Optionally, generating composite sensing data based on the vibration spectrum data and the current ripple fluctuation rate data includes: Performing time domain alignment processing on the vibration spectrum data and the current ripple fluctuation rate data to generate a time domain synchronization feature vector; The energy gradient change feature in the time domain synchronization feature vector is extracted to generate composite sensing data.

[0032] Specifically, time-domain alignment is first performed to generate time-domain synchronized feature vectors. This step presupposes the acquisition of vibration spectrum data and current ripple fluctuation rate data. The vibration spectrum data is first processed into a time-domain vibration sequence. This typically involves analyzing the collected raw vibration time-domain signals or performing an inverse transformation to generate time-series data representing the mechanical vibration state of the equipment. Simultaneously, the acquired current ripple fluctuation rate data undergoes time-domain waveform normalization to generate a current fluctuation time series. Normalization aims to eliminate amplitude variations caused by power supply voltage fluctuations or varying base loads, focusing the data on the fluctuation pattern rather than absolute magnitude, thereby enhancing comparability across different operating conditions. After completing this preprocessing, the time-domain vibration sequence and the current fluctuation time series are precisely aligned on the time axis, ensuring that the vibration value at each time point corresponds to the current fluctuation value at the same instant. This results in a two-dimensional sequence of time-domain synchronized feature vectors, each of which contains information about the mechanical and electrical states at the same instant. Once the time-domain synchronized feature vectors are obtained, the energy gradient variation characteristics are further extracted, which serves as the final composite sensor data. The energy gradient change feature can sensitively capture the moment when a device state undergoes a sudden change or an anomaly begins to emerge. Calculating this feature first defines the instantaneous energy of the time-domain synchronous eigenvector at any moment. This instantaneous energy characterizes the combined intensity of mechanical vibration and current fluctuations at that moment. Subsequently, the temporal gradient of this instantaneous energy—the rate of change of energy over time—is calculated. In a discrete sampling system, the energy gradient can be approximated by calculating the difference between the instantaneous energies at two adjacent sampling moments. This calculation can be expressed as: , in, Representatives in The energy gradient change characteristics at each moment are the desired composite sensing data. for The instantaneous energy at the moment is calculated by the time domain synchronous eigenvector at that moment, which contains The vibration component and current fluctuation component at the moment. Calculated by the following formula: , in for The amplitude of the vibration component at the moment; for The amplitude of the current fluctuation component at the moment; Current fluctuation in the time window The mean within is the weighting coefficient of vibration and current energy, and satisfies , the determination method is: the initial value is set according to the historical fault type ratio of the equipment: when the mechanical fault ratio exceeds 60%, =0.6, =0.4; when the proportion of electrical faults is higher, =0.4, =0.6. Dynamic optimization: under high load Increase by 0.1, run at low speed Increase by 0.1, new equipment running-in period = =0.5. Self-learning calibration: every time 10 fault records are accumulated, if the mechanical fault omission rate exceeds 5%, Increase by 0.05, if the electrical fault omission rate exceeds 5%, Increase by 0.05; is the integration time window, the value is 10 50ms; = is the instantaneous energy at the previous sampling moment. By calculating this energy gradient, the method no longer focuses solely on the absolute values ​​of vibration or current, but instead focuses on the dynamic trend of the combined energy of the two. A dramatic energy gradient change, whether a positive surge or a negative drop, strongly indicates an abnormal transition in the equipment's operating state, such as component impact, sudden load change, or loose connection.

[0033] Optionally, performing time domain alignment processing on the vibration spectrum data and the current ripple fluctuation rate data to generate a time domain synchronization feature vector includes: Based on the vibration spectrum data, a time domain vibration sequence is obtained; Performing time domain waveform normalization processing on the current ripple fluctuation rate data to obtain a current fluctuation time series; The time domain vibration sequence and the current fluctuation time sequence are aligned in the time domain to generate a time domain synchronization feature vector.

[0034] Specifically, first, a time domain vibration sequence is generated based on the acquired vibration spectrum data. Vibration spectrum data describes the vibration energy distribution of the equipment in the frequency domain. In order to use it for time series analysis, it is necessary to use signal processing techniques such as inverse Fourier transform to restore it to a vibration amplitude sequence sampled by time, which is the time domain vibration sequence. , which intuitively reflects the mechanical vibration state of the device at each moment. At the same time, the acquired current ripple fluctuation rate data is subjected to time domain waveform normalization processing to generate a current fluctuation time series. The current ripple fluctuation rate data is a time series that characterizes the severity of high-frequency disturbances in the motor drive current. In order to eliminate the interference caused by the difference in the absolute value of the current when the device is working under different loads, normalization processing must be performed to focus on the shape of the fluctuation rather than the magnitude. An effective normalization method is minimum and maximum value normalization, which is calculated as follows: , In this formula, After normalization, The current fluctuation timing value at the moment. is The original current ripple fluctuation rate data collected at the moment. and They are the maximum and minimum values ​​of the original current ripple fluctuation rate data in a specific monitoring period. Through this processing, the original current data is mapped to a fixed interval, thereby obtaining a current fluctuation time series with consistent form but independent of the absolute load. Finally, the time domain vibration sequence obtained through the above processing is and current fluctuation timing Perform time domain alignment. This step aims to resolve the possible slight deviations in the timestamps of different sensor data acquisitions and ensure that each data point is synchronized in time. By calibrating the timestamps of the two sequences or resampling them to a common time base, it can be ensured that at any sampling moment , we all have a precise corresponding vibration value and current fluctuation value Combining these two values ​​generates the time domain synchronization feature vector Each element of the vector sequence synchronously encapsulates the mechanical and electrical state characteristics of the device at that moment.

[0035] Optionally, the fault tracing analysis includes: Obtaining the conduction threshold and vibration peak timing data of the vibration sensor; Calculating a first propagation delay difference based on the vibration peak time series data; When the first difference in the propagation delay exceeds the conduction threshold, a motor bearing failure is determined.

[0036] Specifically, this method requires two key data points: the first is the conduction threshold of the vibration sensor, which is a key parameter that is pre-set or obtained through experimental calibration.

[0037] The conduction threshold is calibrated through experiments as follows: Sample collection must cover three operating conditions: no load, 50% load, and full load. Under normal conditions, a cumulative data collection of no less than 90 hours is required. Under fault conditions, simulated data of 5 failure stages must be included. The sensor installation position is strictly limited to: the distance between the proximal sensor and the bearing center is ≤5cm, and the horizontal distance between the distal sensor and the proximal sensor is 1.5±0.2m. The threshold calculation adopts a double insurance mechanism, which is based on the normal data. The system uses a rule (covering 99.7% of normal fluctuations) and incorporates the minimum failure threshold of fault data for correction. To accommodate equipment aging, calibration is automatically triggered every 300 hours of operation or after maintenance. The threshold is dynamically updated using a sliding window algorithm to ensure long-term accuracy. This threshold is determined based on the equipment's mechanical structure, material properties, and sensor layout. It represents the time delay criterion used to distinguish whether a fault originates in the motor or elsewhere when a vibration wave propagates between two specific sensors on the equipment. This threshold is determined based on the equipment's mechanical structure, material properties, and sensor layout. The second component is vibration peak time series data, collected by multiple vibration sensors deployed at various locations on the equipment, particularly on the motor, its vicinity, and other key components along the transmission path. By processing the raw vibration signals, impact signals generated by faults such as bearing defects are identified. The time when these impact signal peaks reach each sensor is precisely recorded, forming a time series containing multiple timestamps. Next, the calculation and location process begins. When the first-level diagnostics signal an anomaly, the system analyzes the most recently collected vibration peak time series data. This method uses the peak arrival times of at least two sensors (for example, one close to the motor and another located somewhere on the conveyor rack) to calculate the propagation delay. Based on these timing data, the first difference in propagation delay is calculated as follows: , in, This is the first difference in propagation delay, is the time when the vibration peak reaches the sensor closest to the motor, =The time it takes for the same vibration peak to reach another sensor at a predetermined location. This difference essentially reflects the time difference required for the fault shock wave to propagate from its source to these two different sensors. The second propagation delay difference can be calculated in a similar manner using data from another pair of sensors (such as a motor sensor and a transmission end sensor) to assist in locating or troubleshooting faults in other locations. Finally, a decision is made. The calculated first propagation delay difference is used to The propagation delay difference is compared with the conduction threshold. When the first propagation delay difference exceeds the conduction threshold, the system determines that the fault source is a motor bearing failure. The physical basis of this logic is that if the fault occurs in the motor bearing, the vibration wave is first captured by the sensor near the motor and then propagates along a fixed structural path before being captured by the distant sensor. This generates a specific time delay. The magnitude of this time delay is strongly correlated with the fault source at the motor. The conduction threshold is a critical value set based on this correlation. If the fault occurs elsewhere, the propagation path of the vibration wave will change, causing the calculated propagation delay difference to fall within the threshold range or exhibit another distinguishable pattern.

[0038] Optionally, generating a carton deformation feature based on the laser scanning data includes: Performing dynamic displacement compensation and geometric contour extraction on the laser scanning data to obtain a calibrated three-dimensional edge coordinate set; Deformation physical quantity decoupling calculation is performed on the three-dimensional edge coordinate set to generate carton deformation features.

[0039] Specifically, the system first processes the raw laser scan data, acquired by a laser scanner deployed on the conveyor line. This data is represented as a collection of point clouds describing the spatial positions of the carton surface. Because the carton is dynamic during transport, while the scanner is fixed, the raw point cloud data can experience motion artifacts or distortion. Therefore, dynamic displacement compensation is a crucial first step. This compensation calculates the real-time speed of the conveyor belt and, combined with the timestamp of each point cloud data point, calculates the displacement of that point relative to a fixed reference coordinate system at the instant of scanning. This displacement is then subtracted from the coordinates of the point. This operation transforms the moving carton into a virtual stationary object at the data level, eliminating motion blur and paving the way for accurate geometric analysis. After compensation, geometric contour extraction proceeds. In this step, the system analyzes the compensated, dense point cloud data to identify and extract the key geometric elements that form the carton's basic framework, namely edges and vertices. This is typically achieved through algorithms, such as detecting areas in the point cloud where the normal vector or curvature changes sharply to locate edges, and then determining vertices by calculating the intersection of edge lines. Through this process, the complex and disordered surface point cloud is refined into a set of ordered three-dimensional coordinate points that can represent the macro structure of the carton, that is, the calibrated three-dimensional edge coordinate set. After obtaining this coordinate set, the method enters the core step, which is to perform a decoupling calculation of the deformation physical quantity on this coordinate set to generate the final carton deformation characteristics. The theoretical deformation comparison of the carton load is as follows: Figure 3As shown. The essence of decoupling calculation is to decompose the overall and complex morphological changes of the carton into a series of independent or weakly correlated basic physical deformation components. This calculation compares the measured three-dimensional edge coordinate set with a standard, ideal carton geometric model. Through comparison, multiple dimensions of deformation physical quantities can be calculated. For example, by calculating the deviation between the angle between the measured edges and the ideal 90-degree angle, the shear or twisting amount of the box can be obtained; by calculating the change in the length of the relative edges, the compression or stretching amount of the box can be obtained; by measuring the degree of deviation of the center point of the box surface relative to the plane formed by the vertices, the bulging or concave amount of the box surface can be obtained. These calculated, decoupled physical quantities, such as angle deviation value, edge length change rate, surface curvature, etc., together constitute a multi-dimensional vector or set, which is the deformation feature of the carton.

[0040] Optionally, establishing a dynamic health profile of the device includes: Obtaining current harmonic amplitude ratio data from the equipment operating parameters; Monitoring the current harmonic amplitude ratio data to obtain a time-varying degradation offset; A dynamic health profile of the device is established based on the time-varying degradation offset.

[0041] Specifically, the first step is to obtain data on the current harmonic amplitude ratios from the equipment's operating parameters. Current harmonics refer to components of the motor's operating current that have frequencies that are integer multiples of the fundamental frequency, in addition to the power frequency fundamental. As the motor and its drive system, especially the insulation system, gradually degrade over time due to aging and wear, their nonlinear characteristics intensify, leading to increased distortion of the current waveform and, in turn, slow but continuous changes in the amplitude of specific harmonics. The current harmonic amplitude ratio data is the ratio of the amplitude of a specific harmonic to the amplitude of the fundamental. This ratio effectively eliminates the impact of overall current fluctuations caused by load fluctuations and more purely reflects the changes in the equipment's inherent electrical characteristics. This data is obtained by performing frequency domain analysis methods such as Fourier transform on the real-time motor current signal. Once the current harmonic amplitude ratio data is obtained, the core of the method lies in long-term monitoring and calculation of the time-varying degradation offset. This step first requires collecting current harmonic amplitude ratio data over a period of time when the equipment is in a clear healthy state (for example, after commissioning a new device or after an overhaul), and calculating its stable value as a healthy baseline. Subsequently, during the continuous operation of the equipment, the system will periodically collect this data and compare it with the preset healthy baseline value. The time-varying degradation offset is the difference between the current harmonic amplitude ratio data at the current moment and the healthy baseline value. Its calculation method can be expressed as: , In this formula, Representatives in The time-varying degradation offset calculated at time . is a specific frequency measured at time t ( harmonics) amplitude ratio data. This is the base amplitude ratio of the harmonic in a healthy state. It is not a static value, but a time-varying sequence that quantifies the cumulative degradation of the device's electrical performance since its healthy state. Finally, based on this series of continuously or discretely collected time-varying degradation offsets, a dynamic health profile of the device is established. The dynamic health profile of the device is as follows: Figure 4 As shown in the figure. This file is not a simple status label, but a data set containing timestamps and corresponding degradation offsets. It depicts a degradation trajectory curve of the equipment health status with time as the axis. This dynamic health file records the entire process of the equipment from health to potential failure, and can intuitively show the degradation trend and rate. The degradation of the equipment dynamic health file is as follows: Figure 5 shown.

[0042] Optionally, generating a maintenance decision instruction according to the secondary positioning information includes: Perform fault topology modeling on the secondary location information to obtain a fault location topology map; Maintenance strategy matching and optimization are performed on the fault location topology map to generate maintenance decision instructions.

[0043] Specifically, fault topology modeling is first performed based on the secondary location information derived from fault tracing analysis. This secondary location information identifies the specific faulty component, such as "motor bearing failure." Fault topology modeling does not simply record this name; rather, it places the faulty component within the system structure diagram of the entire carton conveying equipment. This structure diagram is a pre-built digital model, taking the form of a topology graph. Nodes represent physical components within the equipment (such as motors, bearings, drive shafts, tensioners, and conveyor belts), while edges represent the physical connections or functional dependencies between these components. Once the secondary location information is input, the system highlights the corresponding faulty node and its directly associated nodes within the topology graph, generating a fault location topology that clearly illustrates the fault location and its context within the entire system. After obtaining this fault location topology, the system then applies maintenance strategy matching and optimization to it. Maintenance strategy matching involves accessing a pre-built maintenance knowledge base that pre-associates one or more standard maintenance procedures for each component node in the topology graph. For example, the associated strategies for the "motor bearing" node might include "lubrication," "replacement," or "tightening inspection." The system first matches all possible maintenance strategies from the knowledge base based on the nature of the fault (determined by both the primary diagnosis results and the secondary location information). However, matching alone is not enough; the key lies in optimization. This optimization process is based on the fault location topology graph. The system analyzes the connections between nodes in the graph and calculates the weights of different maintenance paths. This means that the system evaluates the sequence of disassembly and installation steps required to execute a particular maintenance strategy, taking into account the time, tools, spare parts costs, and impact on other parts of the equipment. Through this topology-based path analysis and weight calculation, the system compares the comprehensive costs and benefits of different maintenance strategies and selects the optimal maintenance path. For example, for bearing replacement, the system determines a sequence that minimizes the number of disassembled parts and the shortest operation time. The results of this selection and optimization are ultimately integrated into a structured instruction, known as the maintenance decision instruction.

[0044] Optionally, the outputting of a warning signal including a fault component identification and maintenance measures based on the maintenance decision instruction includes: parsing the maintenance decision instruction to obtain optimization strategy parameters for maintenance measures of the faulty component; Adaptive signal encapsulation processing is performed on the optimized strategy parameters of the maintenance measures for the faulty component, and an early warning signal including the faulty component identification and the maintenance measures is output.

[0045] Specifically, the maintenance decision instruction generated in the previous step is first parsed. A maintenance decision instruction is a structured data package containing the optimal maintenance path and strategy. Parsing this instruction involves disassembling the data package and extracting information, resulting in a series of optimized strategy parameters for specific maintenance actions for the faulty component. These parameters are key elements guiding actual maintenance work. For example, they clearly specify the unique identifier or number of the faulty component, the recommended maintenance action (such as "replace" or "lubrication"), the model and specifications of the required spare parts, the detailed sequence of steps for executing the action, as well as the estimated work time and necessary safety precautions. After parsing and obtaining these optimized strategy parameters, the system then performs adaptive signal packaging. This is a critical conversion step, the core of which is "adaptive." This means that the system does not generate a single, fixed alarm format, but instead organizes and packages these optimized strategy parameters in the most appropriate form based on the pre-defined output target or receiver type. For example, if the receiver of the warning signal is a human-machine interface for on-site operators, the encapsulation process will convert the parameters into visual graphic information, which may include highlighting the location of the faulty component on a 3D model of the equipment and presenting the repair steps in a list format. If the receiver is an enterprise-level computerized maintenance management system, the encapsulation process will encode these parameters into data packets that conform to specific communication protocols, such as XML or JSON format, to automatically create repair work orders, request spare parts, and dispatch personnel. If the receiver is a basic audio and visual alarm device, the encapsulation process will generate a simple drive signal to activate a warning light of a specific color and associate it with a pre-recorded voice announcement. The final output of this encapsulation process is a warning signal that includes the faulty component identification and maintenance measures.

[0046] Optionally, performing maintenance strategy matching and optimization on the fault location topology map to generate a maintenance decision instruction includes: Performing node correlation analysis and maintenance path weight calculation on the fault location topology graph to obtain a path weight matrix; Dynamic strategy matching and maintainability index optimization are performed based on the path weight matrix to generate maintenance decision instructions.

[0047] Specifically, we first perform node correlation analysis and maintenance path weight calculation on the fault location topology. In this topology, nodes represent equipment components, and edges represent the connection relationships between components. Node correlation analysis is to identify and quantify these connection relationships. For example, to repair the target fault node, which predecessor nodes need to be disassembled first? There is a strong correlation between these predecessor nodes and the target node. The maintenance path refers to the node sequence that passes through a series of operation steps from the intact state of the equipment to the completion of the repair of the target node. The weight calculation of this path aims to concretize the abstract maintenance difficulty into a numerical value. The weight of each operation step (i.e., an edge in the path) is determined by a combination of multiple factors, and its calculation method can be expressed as: , In this formula, Represents the operation To Operation The combined path weight for this maintenance step. are pre-set importance coefficients, representing the relative importance of time, cost, and risk in decision-making. The importance coefficient is adjusted by combining the hierarchical analysis method with the dynamic scenario. The specific determination method is as follows: the initial value is calculated using the AHP method, and the initial weight ratio of the three is obtained by constructing a judgment matrix of time, cost, and risk. Dynamic adjustment based on equipment operation time: Prioritize efficiency during production peak periods. The value is 0.4 0.5, The value is 0.2 0.3; Prioritize cost control during equipment idle period, The value is 0.4 0.5, α is 0.2 0.3; risk weight Always keep 0.2 The weighting is adjusted to 0.3 to ensure operational safety. Quarterly verification is based on historical data. If the actual impact of a factor deviates from its weight by more than 10%, a fine-tuning of ±0.05 is performed. 、 and They are respectively the normalized values ​​of the estimated time, related costs and operational risks required to complete the step. These raw data are obtained by querying the equipment database and maintenance knowledge base, and are normalized to eliminate the dimensionality effect so that they can be weighted and summed under a unified framework. By performing this calculation for each step on all possible maintenance paths, a path weight matrix can be constructed, which fully describes the comprehensive "cost" of performing different maintenance operations. After obtaining the path weight matrix, the system performs dynamic strategy matching and maintainability index optimization based on the matrix. The maintainability index is a comprehensive indicator for evaluating the pros and cons of a maintenance plan. It is defined as a quantitative value that combines the total weight of the path, the scope of fault impact and the correlation between components. The calculation formula is:

[0048] in, is the total path weight, is the fault impact range coefficient (0-1, the greater the impact, the closer the value is to 1), is the component correlation correction coefficient (0.1-0.3, the higher the correlation, the larger the value), and are the number of normally functioning components after maintenance and the total number of components in the equipment, respectively. The maintainability index is inversely proportional to the total path weight. The lower the total weight, the smaller the fault impact, and the higher the component correlation, the higher the maintainability index and the better the solution. Dynamic strategy matching involves the system accessing all feasible maintenance strategies (such as "replace," "repair," and "adjust") from a knowledge base for the faulty node. Using the path weight matrix, the system calculates the optimal execution path for each strategy—the path with the lowest total weight. This is a typical graph theory shortest path optimization problem. The maintainability index is a comprehensive metric used to evaluate the quality of a maintenance solution. It is generally inversely proportional to the total path weight. The lower the total weight, the higher the maintainability index and the better the solution. The optimization process involves using algorithms (such as Dijkstra or A*) to traverse all possible strategies and their optimal paths, searching for the combination that maximizes the maintainability index. The strategy and its execution path ultimately selected with the highest maintainability index constitute the final maintenance decision.

[0049] Based on the same inventive concept, Figure 6 As shown, the present invention also provides a self-diagnosis and early warning system for the operation status of a carton conveying equipment, the system comprising: Multi-source signal acquisition module, used to obtain vibration spectrum data, current ripple fluctuation rate data, equipment operating parameters, vibration wave transmission time series data and laser scanning data; A composite sensing generation module, configured to generate composite sensing data based on the vibration spectrum data and the current ripple fluctuation rate data; A deformation feature extraction module, configured to generate carton deformation features based on the laser scanning data; A health record modeling module, configured to establish a dynamic health record of the device based on the device operating parameters; an abnormality diagnosis module, configured to perform abnormality separation processing on the composite sensor data and the dynamic health record of the device to generate a primary diagnosis result; A fault tracing module is used to perform fault tracing analysis on the vibration wave transmission time series data and the carton deformation characteristics based on the first-level diagnosis results to generate second-level positioning information; A maintenance decision module, configured to generate a maintenance decision instruction based on the secondary positioning information; The early warning output module is used to output an early warning signal including a fault component identification and maintenance measures based on the maintenance decision instruction.

[0050] To verify the feasibility of this invention, it was applied to a large automated logistics center. This center has dozens of high-speed carton conveyor lines used for sorting and transferring packages. Traditional maintenance methods rely on regular inspections and post-event repairs. When hidden equipment failures, such as roller jams and bearing wear, occur, they are difficult to quickly locate, often leading to unplanned downtime of the conveyor lines and impacting overall sorting efficiency. The logistics center hopes to use the method of this invention to implement real-time self-diagnosis and predictive maintenance of the carton conveyor equipment.

[0051] In this embodiment, the logistics center deployed the system of the present invention on a key conveyor line. Using a multi-source signal acquisition module, the system installed vibration and current sensors on the conveyor line's drive motor, key bearing locations, and frame. A laser scanner was also installed above the conveyor line's end. This system continuously acquired vibration spectrum data, current ripple fluctuation rate data, vibration wave transmission time series data, laser scanning data, and equipment operating parameters, including the motor current harmonic amplitude ratio.

[0052] The system first generates composite sensor data that is highly sensitive to combined mechanical and electrical disturbances through time-domain alignment and energy gradient feature extraction based on acquired vibration spectrum data and current ripple fluctuation rate data. Furthermore, based on long-term monitoring of current harmonic amplitude ratio data, a unique dynamic health profile for the conveyor line is established, documenting performance degradation offsets from new installation to several months of operation.

[0053] At 3:20 PM on August 15, 2023, the system's abnormality diagnosis module detected persistent small spikes in the composite sensor data. This characteristic deviated significantly from the normal operating mode in the equipment's dynamic health record, and immediately generated a first-level diagnostic result, determining that the equipment had an early potential fault. After receiving the first-level diagnostic result, the fault tracing module was activated. This module analyzed the vibration wave transmission timing data and calculated that the first difference in propagation delay from the sensor near the motor to the sensor in the middle of the frame was 3.5ms. This value exceeded the preset 2.0ms transmission threshold and initially pointed to the source of the fault as the drive motor. At the same time, the deformation feature extraction module analyzed the laser scanning data of the cartons passing through this section and found slight, regular scratches on the bottom of several cartons. After decoupling the deformation physical quantity, this feature was highly consistent with the small high-frequency vibration model caused by early wear of the motor bearing. Combining these two pieces of information, the system generated second-level positioning information: the output end bearing of drive motor A03 was showing early wear.

[0054] Based on this secondary location information, the maintenance decision module constructed a fault location topology map encompassing motor A03, its coupling, gearbox, and its base. While matching the standard bearing replacement maintenance strategy, the system, through node correlation analysis and querying the equipment's dynamic health records, discovered that the coupling associated with this bearing had reached 85% of its wear life. To avoid a second downtime in the short term, the system optimized the maintainability index and generated a maintenance decision instruction: "Replace motor A03's bearing and coupling at once." Finally, the early warning output module packaged this instruction into an early warning signal containing the following: "Faulty component identification: motor A03 bearing," "Maintenance action: Replace," and "Recommended coordinated maintenance: Replace coupling." This was then sent to the maintenance team's mobile device and automatically generated a requisition form in the spare parts system. Following the precise instructions, maintenance personnel completed the task within the nighttime maintenance window in just 45 minutes, avoiding a potential daytime production interruption.

[0055] Data shows that the system, implemented in this pilot project, enables the logistics center to accurately predict and efficiently maintain conveyor line faults. Compared to traditional troubleshooting methods that rely on manual experience, this system reduces fault location time from an average of 2-3 hours to less than 5 minutes, achieving a diagnostic accuracy rate exceeding 98%. Through predictive maintenance, the pilot conveyor line's average monthly unplanned downtime has been reduced by 70%, and spare parts costs have been reduced by approximately 15% by avoiding misidentification and replacement of non-faulty components.

[0056] Table 1 Comparison of fault diagnosis efficiency of carton conveying equipment

[0057] Table 2 Comparison of equipment maintenance and operation benefit data

[0058] Table 3. Example data sheet for diagnosis and decision-making of specific fault events (bearing wear of motor A03)

[0059] It can be seen from the data in Tables 1 to 3 above that the present invention has achieved remarkable technical effects in the practical application of carton conveying equipment. Table 1 clearly demonstrates the huge advantages of the present invention in fault diagnosis efficiency, especially shortening the fault location time of minutes by more than 97%, and being able to provide early warning, which fundamentally changes the maintenance mode. The data in Table 2 proves that the improvement in diagnostic efficiency is directly converted into considerable economic and operational benefits. The significant reduction in unplanned downtime and maintenance costs directly improves the operational efficiency of the logistics center. Table 3 records in detail the entire process of a real fault from detection to decision-making, demonstrating the logical rigor and data-driven characteristics of the method of the present invention. Through mutual verification of multi-dimensional information, the accuracy of positioning and the optimization of decision-making are ensured, which fully proves the advancement and practicality of the present invention in realizing intelligent maintenance of equipment.

[0060] It should be noted that the electrical connections between the above-mentioned units do not necessarily mean direct connections of lines. Indirect connections are applicable to the embodiments of the present invention as long as the purpose of the present invention is achieved. The above description is only an exemplary embodiment of the present invention and is not intended to limit the scope of the present invention.

[0061] That is, any equivalent changes and modifications made according to the teachings of the present invention are still within the scope of the present invention. Those skilled in the art will readily conceive of other embodiments of the present invention after considering the disclosure of the specification and practical truths. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary technical means in the art not described herein.

Claims

1. A method for self-diagnosis and early warning of the operating status of carton conveying equipment, characterized in that: The method comprises: Obtain vibration spectrum data, current ripple fluctuation rate data, equipment operating parameters, vibration wave transmission timing data and laser scanning data; generating composite sensing data based on the vibration spectrum data and the current ripple fluctuation rate data; generating a carton deformation feature based on the laser scanning data; Establishing a dynamic health profile of the equipment based on the equipment operating parameters; Performing abnormality separation processing on the composite sensor data and the dynamic health record of the device to generate a primary diagnosis result; Based on the first-level diagnostic results, a fault tracing analysis is performed on the vibration wave transmission time series data and the carton deformation characteristics to generate second-level positioning information; generating maintenance decision instructions based on the secondary positioning information; Based on the maintenance decision instruction, an early warning signal including a fault component identification and maintenance measures is output.

2. A method for self-diagnosis and early warning of the operation status of carton conveying equipment according to claim 1, characterized in that: The generating of composite sensing data based on the vibration spectrum data and the current ripple fluctuation rate data includes: Performing time domain alignment processing on the vibration spectrum data and the current ripple fluctuation rate data to generate a time domain synchronization feature vector; The energy gradient change feature in the time domain synchronization feature vector is extracted to generate composite sensing data.

3. A method for self-diagnosis and early warning of the operating status of carton conveying equipment according to claim 2, characterized in that: The performing time domain alignment processing on the vibration spectrum data and the current ripple fluctuation rate data to generate a time domain synchronization feature vector includes: Based on the vibration spectrum data, a time domain vibration sequence is obtained; Performing time domain waveform normalization processing on the current ripple fluctuation rate data to obtain a current fluctuation time series; The time domain vibration sequence and the current fluctuation time sequence are aligned in the time domain to generate a time domain synchronization feature vector.

4. The method for self-diagnosis and early warning of the operating status of carton conveying equipment according to claim 1, characterized in that: The fault tracing analysis includes: Obtaining the conduction threshold and vibration peak timing data of the vibration sensor; Calculating a first propagation delay difference based on the vibration peak time series data; When the first difference in the propagation delay exceeds the conduction threshold, a motor bearing failure is determined.

5. The method for self-diagnosis and early warning of the operating status of carton conveying equipment according to claim 1 is characterized in that: Generating the carton deformation feature based on the laser scanning data includes: Performing dynamic displacement compensation and geometric contour extraction on the laser scanning data to obtain a calibrated three-dimensional edge coordinate set; Deformation physical quantity decoupling calculation is performed on the three-dimensional edge coordinate set to generate carton deformation features.

6. The method for self-diagnosis and early warning of the operating status of carton conveying equipment according to claim 1, characterized in that: The establishment of a dynamic health profile of the device includes: Obtaining current harmonic amplitude ratio data from the equipment operating parameters; Monitoring the current harmonic amplitude ratio data to obtain a time-varying degradation offset; A dynamic health profile of the device is established based on the time-varying degradation offset.

7. The method for self-diagnosis and early warning of the operation status of carton conveying equipment according to claim 1 is characterized in that: Generating a maintenance decision instruction according to the secondary positioning information includes: Perform fault topology modeling on the secondary location information to obtain a fault location topology map; Maintenance strategy matching and optimization are performed on the fault location topology map to generate maintenance decision instructions.

8. The method for self-diagnosis and early warning of the operation status of carton conveying equipment according to claim 1 is characterized in that: Outputting a warning signal including a fault component identification and maintenance measures based on the maintenance decision instruction includes: parsing the maintenance decision instruction to obtain optimization strategy parameters for maintenance measures of the faulty component; Adaptive signal encapsulation processing is performed on the optimized strategy parameters of the maintenance measures for the faulty component, and an early warning signal including the faulty component identification and the maintenance measures is output.

9. The method for self-diagnosis and early warning of the operation status of carton conveying equipment according to claim 7, characterized in that: The performing maintenance strategy matching and optimization on the fault location topology map to generate a maintenance decision instruction includes: Performing node correlation analysis and maintenance path weight calculation on the fault location topology graph to obtain a path weight matrix; Dynamic strategy matching and maintainability index optimization are performed based on the path weight matrix to generate maintenance decision instructions.

10. A carton conveying equipment operating status self-diagnosis and early warning system, applied to a carton conveying equipment operating status self-diagnosis and early warning method according to any one of claims 1 to 9, characterized in that: The system comprises: Multi-source signal acquisition module, used to obtain vibration spectrum data, current ripple fluctuation rate data, equipment operating parameters, vibration wave transmission time series data and laser scanning data; A composite sensing generation module, configured to generate composite sensing data based on the vibration spectrum data and the current ripple fluctuation rate data; A deformation feature extraction module, configured to generate carton deformation features based on the laser scanning data; A health record modeling module, configured to establish a dynamic health record of the device based on the device operating parameters; an abnormality diagnosis module, configured to perform abnormality separation processing on the composite sensor data and the dynamic health record of the device to generate a primary diagnosis result; A fault tracing module is used to perform fault tracing analysis on the vibration wave transmission time series data and the carton deformation characteristics based on the first-level diagnosis results to generate second-level positioning information; A maintenance decision module, configured to generate a maintenance decision instruction based on the secondary positioning information; The early warning output module is used to output an early warning signal including a fault component identification and maintenance measures based on the maintenance decision instruction.

Images