Real-time data-driven roll-over machine hooking robot failure analysis method

Abstract

This invention discloses a real-time data-driven fault analysis method for tipper unhooking robots, relating to the field of fault analysis for tipper unhooking robots. The method includes: identifying multiple functional units; establishing normal collaborative relationships and a set of abnormal collaborative influence relationships when any one or more units malfunction; performing fault feature discrimination analysis to establish a combination of M functional units whose discrimination meets a preset discrimination index; constructing a fault detection model using the abnormal collaborative influence relationships and normal collaborative relationships corresponding to each functional unit in the M functional unit combinations; collecting real-time operating data and inputting it into the fault detection model for fault analysis, and outputting a fault warning signal. This method solves the technical problems of difficulty in fault feature extraction, complex model construction, and difficulty in achieving real-time warning and rapid response in existing fault analysis methods, achieving the technical effects of improving feature extraction accuracy, reducing model construction complexity, and realizing real-time warning and rapid response.

Description

Technical Field

[0001] This application relates to the field of fault analysis of tipper unhooking robots, and in particular to a real-time data-driven fault analysis method for tipper unhooking robots. Background Technology

[0002] As a key piece of equipment in tippler systems, the uncoupling robot is responsible for automatically completing the uncoupling operation of wagons during the tipping process in the transportation of bulk cargo such as coal and ore on railways, greatly improving operational efficiency and safety. However, various malfunctions may occur during the operation of the uncoupling robot. These malfunctions can not only affect the normal operation of the tippler but may also lead to equipment damage or even safety accidents. In the field of fault analysis for tippler uncoupling robots, existing data-driven fault diagnosis requires the preprocessing and feature extraction of large amounts of data and the construction of complex mathematical models. This often results in fault analysis results lagging behind the actual occurrence of the fault, failing to achieve real-time early warning and rapid response. This not only misses the optimal fault handling opportunity but may also exacerbate the severity of the fault and expand its impact.

[0003] At present, the fault analysis of tipper unhooking robots faces technical problems such as difficulty in extracting fault features and complexity in model construction, which makes it difficult to achieve real-time fault warning and rapid response. Summary of the Invention

[0004] This application provides a real-time data-driven fault analysis method for a tipper unhooking robot. It employs a method that identifies multiple functional units of the robot, establishes sets of normal and abnormal collaborative influence relationships, performs fault feature discrimination analysis, and constructs M functional unit combinations with discrimination scores that meet preset indicators. Based on these functional unit combinations and their corresponding abnormal and normal collaborative influence relationships, a fault detection model is built. Real-time operational data is collected and input into the model for fault analysis, and fault warning signals are output. These techniques achieve the technical effects of improving the accuracy of fault feature extraction, reducing model construction complexity, and realizing real-time early warning and rapid response.

[0005] This application provides a real-time data-driven fault analysis method for a tipper unhooking robot, comprising: identifying multiple functional units of the tipper unhooking robot; establishing normal collaborative relationships among the multiple functional units, and a set of abnormal collaborative influence relationships when any one or more units malfunction; performing fault feature discrimination analysis using the set of abnormal collaborative influence relationships, and establishing M combinations of functional units whose discrimination satisfies a preset discrimination index, where M is an integer greater than 1; constructing a fault detection model using the abnormal collaborative influence relationships corresponding to each functional unit in the M combinations of functional units and the normal collaborative relationships; collecting real-time operating data of the multiple functional units and inputting it into the fault detection model for fault analysis, and outputting a fault warning signal.

[0006] In a possible implementation, the set of abnormal collaborative influence relationships is used to perform a fault feature discrimination analysis to establish M functional unit combinations whose discrimination satisfies a preset discrimination index. The following processing is performed: First abnormal collaborative influence relationship and second abnormal collaborative influence relationship are extracted from the set of abnormal collaborative influence relationships, along with corresponding first abnormal unit labeling information and second abnormal unit labeling information. Here, abnormal collaborative influence relationship refers to the collaborative relationship exhibited by two or more functional units under abnormal conditions, and the first abnormal collaborative influence relationship and second abnormal collaborative influence relationship refer to any two different abnormal collaborative influence relationships in the set of abnormal collaborative influence relationships. Discrimination analysis for abnormal identification is performed on the collaborative relationship data between functional units in the first abnormal collaborative influence relationship and the second abnormal collaborative influence relationship to generate a first discrimination index. If the first discrimination index is less than the preset discrimination index, the labeled abnormal units in the first abnormal unit labeling information and the second abnormal unit labeling information are merged into a single functional unit combination and added to the M functional unit combinations. If the first discrimination index is greater than or equal to the preset discrimination index, the labeled abnormal units in the first abnormal unit labeling information and the second abnormal unit labeling information are each added as a functional unit combination to the M functional unit combinations.

[0007] In a possible implementation, the determination of the preset discrimination index involves the following steps: determining a preset network structure for constructing the fault detection model; configuring a binary classifier with the preset network structure and constructing several sets of collaborative relationship sample data corresponding to several data discriminations; testing the binary classifier with the several sets of collaborative relationship sample data, obtaining the minimum discrimination with a test classification accuracy greater than a preset threshold, and setting it as the preset discrimination index.

[0008] In a possible implementation, a fault detection model is constructed using the abnormal collaborative influence relationships and normal collaborative relationships corresponding to each functional unit in the M functional unit combination, and the following processing is performed: topological transformation is performed on each abnormal collaborative influence relationship and normal collaborative relationship corresponding to each functional unit in the M functional unit combination to generate each abnormal collaborative influence topology and normal collaborative relationship topology; each topological difference vector between each abnormal collaborative influence topology and the normal collaborative relationship topology is calculated; and the fault detection model is trained using each topological difference vector.

[0009] In a possible implementation, after outputting the fault warning signal, the following processing is performed: determining whether the fault warning signal includes warnings for multiple faulty functional units; if so, connecting to the fault log database to collect multiple historical fault record datasets for the multiple faulty functional units; performing fault probability analysis on the multiple historical fault record datasets at the current moment to generate multiple fault probabilities corresponding to the multiple faulty functional units; and arranging the multiple fault probabilities in descending order to generate a recommended maintenance priority sequence, which is sent to the control terminal of the tipper unhooking robot together with the fault warning signal.

[0010] In a possible implementation, the failure probability at the current moment is analyzed using the multiple historical failure record datasets to generate multiple failure probabilities corresponding to the multiple failure functional units. The following processing is performed: based on the multiple historical failure record datasets, the failure frequency change analysis is performed on the multiple failure functional units to generate multiple failure frequency features; the difference between the next failure time and the current moment is predicted using the multiple failure frequency features to generate multiple first failure probabilities; real-time operation data of the tipper unhooking robot is collected, and the operation scenario and load characteristics are analyzed to analyze the failure probability of the multiple failure functional units under the current operation scenario and load characteristics, and the multiple first failure probabilities are compensated to generate the multiple failure probabilities.

[0011] In a possible implementation, the failure probability of the multiple faulty functional units under the current work scenario and load characteristics is analyzed, and the following processing is performed: the frequency distribution characteristics of the faulty work scenario and faulty load characteristics of the multiple faulty functional units are analyzed based on the multiple historical fault record datasets; based on the frequency distribution characteristics, the proportion of the corresponding distribution frequency of the work scenario and load characteristics to the total fault frequency in the frequency distribution characteristics is calculated, and the failure probability of the multiple faulty functional units under the current work scenario and load characteristics is generated.

[0012] In a possible implementation, the following processing is performed: the plurality of functional units include at least the robotic arm, gripping device, vision system, control system, and drive system of the tipper unhooking robot.

[0013] This application proposes a real-time data-driven fault analysis method for a tipper unhooking robot. First, it identifies multiple functional units of the tipper unhooking robot. Then, it establishes the normal collaborative relationships among these functional units, as well as a set of abnormal collaborative influence relationships when any one or more units malfunction. Next, it performs a fault feature discrimination analysis using the set of abnormal collaborative influence relationships, establishing M combinations of functional units (M being an integer greater than 1) whose discrimination satisfies a preset discrimination index. Then, it constructs a fault detection model using the abnormal collaborative influence relationships corresponding to each functional unit in the M combinations and the normal collaborative relationships. Finally, it collects real-time operating data from the multiple functional units and inputs it into the fault detection model for fault analysis, outputting a fault warning signal. This method achieves the technical effects of improving the accuracy of fault feature extraction, reducing model construction complexity, and realizing real-time warning and rapid response. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the method according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0015] Figure 1 This is a flowchart illustrating the real-time data-driven fault analysis method for a tipper unhooking robot provided in an embodiment of this application.

[0016] Figure 2 This is a flowchart illustrating the process of generating a recommended maintenance priority sequence in the real-time data-driven fault analysis method for a tipper unhooking robot provided in this application embodiment. Detailed Implementation

[0017] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below.

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0020] This application provides a real-time data-driven fault analysis method for a tipper unhooking robot, such as... Figure 1 As shown, the method includes:

[0021] Step S100: Determine the multiple functional units of the tipper unhooking robot.

[0022] Specifically, a thorough understanding of the overall structure and working principle of the tippler unhooking robot is required. This involves analyzing the robot's design drawings, instruction manuals, or existing system architecture diagrams to automatically identify or manually designate multiple independent but interconnected functional units. These functional units refer to the components or systems within the robot that perform specific functions or tasks, and can include the robot's robotic arm, sensors, control system, power system, etc.

[0023] In one possible implementation, step S100 further includes step S110, wherein the plurality of functional units include at least the robotic arm, gripping device, vision system, control system, and drive system of the tipper unhooking robot.

[0024] Specifically, the robot's multiple functional units include at least a robotic arm, a gripping device, a vision system, a control system, and a drive system. Among these, the robotic arm is the main actuator of the tipper unhooking robot, responsible for performing the unhooking action. Its design includes multiple joints and actuators to achieve flexible movement. Determining the robotic arm as one of the functional units requires clarifying its key parameters such as range of motion, speed, and load capacity through simulation or actual testing, and analyzing its motion characteristics under normal and abnormal conditions.

[0025] The gripping device is the end effector of a robotic arm, used to grasp and release goods. Its design includes different forms such as grippers and suction cups. To determine whether the gripping device is a functional unit, it is necessary to evaluate its gripping force, gripping accuracy, adaptability, and other performance characteristics through actual testing, and analyze its performance under different working conditions.

[0026] The vision system is used to identify information such as the position, shape, and size of goods, providing a basis for robot positioning and control. It includes cameras and image processing algorithms. To define the vision system as a functional unit, its recognition accuracy, speed, and stability need to be tested under different lighting conditions, angles, and occlusions. Based on the test results, the image processing algorithm of the vision system should be optimized to improve recognition speed and accuracy.

[0027] The control system is the brain of a robot, responsible for receiving input signals, processing data, and issuing control commands. It includes components such as a PLC (Programmable Logic Controller) and a computer. Defining the control system as a functional unit requires analyzing its control algorithm, response time, stability, and other performance characteristics, as well as its control strategies under different operating conditions.

[0028] The drive system provides power to the robot and includes motors, reducers, and transmission devices. Defining the drive system as a functional unit requires testing its driving force, efficiency, reliability, and other performance characteristics, as well as analyzing its performance under different loads and operating conditions. This approach, by refining the functional units, allows for a better understanding of the robot's working principles and failure modes. When building a fault detection model, it allows for more precise selection of input features and model structure, thereby improving the model's predictive performance and robustness, and achieving more accurate and efficient fault detection and early warning.

[0029] Step S200: Establish the normal collaborative relationship of the multiple functional units, and the set of abnormal collaborative impact relationships when any one or more units are abnormal.

[0030] Specifically, after identifying the functional units, the analysis focuses on how these units collaborate normally and how other units are affected when one or more units malfunction. This can be achieved through simulation and historical data analysis. For example, a collaboration relationship database can be established to store the normal collaboration rules and abnormal collaboration impact relationships between functional units. These rules can be learned from historical data using machine learning algorithms.

[0031] Step S300: Perform a fault feature discrimination analysis using the set of abnormal collaborative influence relationships, and establish a combination of M functional units whose discrimination satisfies the preset discrimination index, where M is an integer greater than 1.

[0032] Specifically, data in the collaborative relationship database is mined to calculate the discriminative power indices (SPIs) for different functional unit combinations during failures, such as information gain and the Gini coefficient. Then, M functional unit combinations that meet the preset SPIs are selected. The SPIs measure the ability of a functional unit combination to distinguish between different types of failures. The preset SPIs are thresholds used to determine whether abnormal collaborative influence relationships have sufficient discriminative power.

[0033] In one possible implementation, the set of abnormal collaborative influence relationships is used to perform a distinguishability analysis of fault characteristics, establishing M functional unit combinations whose distinguishability meets a preset distinguishability index. Step S300 further includes step S310, extracting the first and second abnormal collaborative influence relationships from the set of abnormal collaborative influence relationships, along with the corresponding first and second abnormal unit labeling information. Here, an abnormal collaborative influence relationship refers to the collaborative relationship exhibited by two or more functional units under abnormal conditions, and the first and second abnormal collaborative influence relationships refer to any two distinct abnormal collaborative influence relationships in the set of abnormal collaborative influence relationships. Specifically, the set of abnormal collaborative influence relationships is traversed to extract specific pairs of abnormal collaborative influence relationships. An abnormal collaborative influence relationship refers to the collaborative relationship exhibited by two or more functional units under abnormal conditions, revealing the propagation path or common cause of the fault. Each pair of abnormal collaborative influence relationships contains abnormal interaction information between at least two functional units. Simultaneously, the first and second abnormal unit labeling information corresponding to these abnormal collaborative influence relationships are extracted. This information is used to identify which functional units participated in the abnormal collaboration and includes data used to identify the abnormal state of the functional units, such as attributes like abnormality type, timestamp, and severity.

[0034] Step S320: Perform anomaly identification discrimination analysis on the collaboration relationship data between functional units in the first and second abnormal collaboration relationships to generate a first discrimination score. Specifically, preprocessing operations such as cleaning and normalization are performed on the extracted collaboration relationship data to improve the accuracy of the analysis. The first discrimination score, i.e., the effectiveness of abnormal collaboration relationships in distinguishing different fault characteristics, is calculated using statistical indicators (such as mutual information and correlation coefficients) or machine learning models (such as classifiers and clustering algorithms).

[0035] Step S330: If the first distinguishability is less than the preset distinguishability index, the marked abnormal units in the first and second abnormal unit marking information are merged into a functional unit combination and added to the M functional unit combinations. Specifically, when the first distinguishability is less than the preset distinguishability index, it is considered that the currently analyzed abnormal collaborative influence relationship is not effective enough in distinguishing fault characteristics. Therefore, according to the abnormal unit marking information, units with similar abnormal characteristics or close collaborative relationships are merged to form a new functional unit combination. The merged functional unit combination is added to the M functional unit combinations, and the relevant records are updated simultaneously.

[0036] Step S340: If the first discrimination index is greater than or equal to the preset discrimination index, the marked abnormal units in the first and second abnormal unit marking information are each added as a functional unit combination to the M functional unit combinations. Specifically, when the first discrimination index is greater than or equal to the preset discrimination index, the currently analyzed abnormal synergistic influence relationship is considered sufficiently effective in distinguishing fault features. Therefore, based on the abnormal unit marking information, each abnormal unit is treated as an independent functional unit combination. These independent functional unit combinations are added to the M functional unit combinations. This implementation optimizes the construction process of functional unit combinations through detailed discrimination analysis, improving the accuracy and efficiency of the fault detection model.

[0037] In one possible implementation, the determination of the preset discrimination index, step S330 further includes step S331, determining the preset network structure for constructing the fault detection model. Specifically, based on the working characteristics and fault detection requirements of the tipper unhooking robot, a suitable neural network structure is selected, such as a convolutional neural network (CNN), a recurrent neural network (RNN), or a deep neural network (DNN). The number of network layers, nodes, activation functions, and other parameters are configured to ensure that the network can handle the scale and complexity of the input data.

[0038] Step S332: Configure a binary classifier using the preset network structure and construct several sets of collaborative relationship sample data corresponding to several data discrimination levels. Specifically, based on the preset network structure, add an output layer and configure it as a binary classification task to distinguish whether abnormal collaborative influence relationships are effective or ineffective in distinguishing fault features. Based on the historical operation data and fault records of the tipper unhooking robot, construct several sets of collaborative relationship sample data. Each set of data includes collaborative relationship data between functional units (such as sensor readings, control signals, etc.) and corresponding labels (effective or ineffective).

[0039] Step S333: The binary classifier is tested using the aforementioned sets of collaborative relationship sample data to obtain the minimum discriminant value at which the test classification accuracy exceeds a preset threshold, which is then set as the preset discriminant value index. Specifically, the constructed collaborative relationship sample data is used to test the binary classifier and evaluate its classification performance. By adjusting the discriminant value, the minimum discriminant value at which the test classification accuracy exceeds a preset threshold (such as 90%, 95%, etc.) is found and set as the preset discriminant value index. This implementation method, by adjusting the discriminant value and testing classification performance, can find the feature threshold that best distinguishes between effective and ineffective abnormal collaborative influence relationships, thereby improving the accuracy of the model.

[0040] Step S400: Construct a fault detection model based on the abnormal collaborative influence relationship and the normal collaborative relationship corresponding to each functional unit in the M functional unit combination.

[0041] Specifically, after identifying functional unit combinations with significant distinguishability, these combinations can be used to construct a fault detection model. This model can be rule-based, statistical, or machine learning-based. Based on functional unit combinations, abnormal collaborative influence relationships, and normal collaborative relationships, a machine learning model is trained. The model's goal is to determine the presence of faults and predict fault types based on real-time operational data.

[0042] In one possible implementation, a fault detection model is constructed using the abnormal collaborative influence relationships and normal collaborative relationships corresponding to each functional unit in the M functional unit combination. Step S400 further includes step S410, which performs topological transformations on the abnormal collaborative influence relationships and normal collaborative relationships corresponding to each functional unit in the M functional unit combination, generating abnormal collaborative influence topologies and normal collaborative relationship topologies. Specifically, using topological representation methods in graph theory, each functional unit in the M functional unit combination is defined as a node and assigned a unique identifier. Connections (edges) between nodes are defined according to the collaborative relationships between functional units. Normal collaborative relationships are represented by solid lines, and abnormal collaborative influence relationships are represented by dashed lines or lines of different colors. Based on the definitions of nodes and edges, normal collaborative relationship topologies and abnormal collaborative influence topologies are constructed. These topological structures can be undirected or directed graphs, depending on the nature of the collaborative relationships.

[0043] Step S420: Calculate the topological difference vectors between each of the anomalous collaborative influence topologies and the normal collaborative relationship topologies. Specifically, using topological difference analysis, for each anomalous collaborative influence topology, calculate its difference from the normal collaborative relationship topology by comparing attributes such as the position of nodes, the number of edges, and the weight of edges in the two topologies. Represent the calculated differences in vector form, where the magnitude of the vector indicates the degree of difference, and the direction of the vector indicates the type or direction of the difference (e.g., increase, decrease, change, etc.).

[0044] Step S430: Train the fault detection model using the various topological difference vectors. Specifically, machine learning or deep learning methods are employed to construct a training set using the calculated topological difference vectors as training data. The training set is used to train the model (e.g., a neural network), enabling the model to learn the difference between normal cooperative relationships and abnormal cooperative effects. The model's performance is evaluated through cross-validation or a test set to ensure that the model can accurately identify functional units with potential anomalies. This implementation quantifies the difference between normal cooperative relationships and abnormal cooperative effects by calculating topological difference vectors, providing a reliable basis for fault detection and improving the accuracy and efficiency of fault analysis.

[0045] Step S500: Collect real-time operating data of the multiple functional units and input them into the fault detection model for fault analysis, and output fault warning signals.

[0046] Specifically, in practical applications, real-time operational data (such as sensor readings and system logs) of the tipper unhooking robot is continuously collected through sensor networks and data interfaces. This data is then preprocessed and input into a fault detection model for analysis. The model outputs a prediction result indicating whether a fault exists and its type. If a fault exists, an early warning mechanism is triggered, outputting a fault warning signal. This embodiment of the application employs multiple functional units of the robot, establishes sets of normal and abnormal collaborative influence relationships, performs fault feature discrimination analysis, and constructs M functional unit combinations with discrimination scores meeting preset indicators. Based on these functional unit combinations and their corresponding abnormal and normal collaborative influence relationships, a fault detection model is constructed. Real-time operational data is collected and input into the model for fault analysis, and fault warning signals are output. These techniques improve the accuracy of fault feature extraction, reduce model construction complexity, and achieve real-time early warning and rapid response.

[0047] like Figure 2 As shown, in one possible implementation, after outputting the fault warning signal, the method further includes step S600, determining whether the fault warning signal contains warnings for multiple faulty functional units. Specifically, the warning signal is parsed to identify the specific functional unit it points to, and it is determined whether there are multiple such units.

[0048] Step S700: If yes, connect to the fault log database to collect multiple historical fault record datasets of the multiple faulty functional units. Specifically, once it is determined that the warning signal involves multiple functional units, the historical fault records of these functional units are obtained by connecting to the fault log database (a database storing historical fault records of each functional unit of the tipper unhooking robot). For each faulty functional unit, its historical fault records are retrieved to form multiple historical fault record datasets.

[0049] Step S800: Analyze the failure probability at the current moment using the multiple historical fault record datasets to generate multiple failure probabilities corresponding to the multiple faulty functional units. Specifically, machine learning or statistical methods are used to analyze the historical fault record datasets and estimate the probability of each faulty functional unit failing at the current moment. For example, information such as historical fault frequency, fault type, and environmental conditions in which the fault occurred can be used to train a prediction model, which then generates a failure probability value for each faulty functional unit.

[0050] In step S900, based on the multiple fault probabilities, a recommended maintenance priority sequence is generated by arranging them in descending order. This sequence, along with the fault warning signal, is sent to the control terminal of the tippler unhooking robot. Specifically, the fault probability values ​​generated in step 800 are sorted in descending order to form a recommended maintenance priority sequence. This sequence indicates which functional units are most likely to fail and therefore should be prioritized for maintenance. Finally, this priority sequence, along with the original fault warning signal, is sent to the control terminal of the tippler unhooking robot (the robot's operating interface or control system, used to receive fault warnings and maintenance suggestions) so that the operator or automation system can take appropriate maintenance measures. This implementation, by providing a maintenance priority sequence based on fault probabilities, can more effectively allocate maintenance resources and reduce unnecessary downtime.

[0051] In one possible implementation, the failure probability at the current moment is analyzed using the multiple historical fault record datasets to generate multiple failure probabilities corresponding to the multiple fault functional units. Step S800 further includes step S810, which analyzes the change in failure frequency for each of the multiple fault functional units based on the multiple historical fault record datasets to generate multiple failure frequency features. Specifically, statistical analysis and data mining methods are used to process the multiple historical fault record datasets. For each fault functional unit, its failure frequency (the number of times a failure occurs per unit time) in different time periods is calculated by plotting time series graphs, calculating moving averages, or applying time series analysis models (such as the ARIMA model), and the trend of these frequencies over time is observed. Fault frequency features may include the average value, standard deviation, maximum / minimum value, and trend (increasing / decreasing / stable) of the failure frequency.

[0052] Step S820 involves predicting the difference between the next fault time and the current time based on the multiple fault frequency features, generating multiple first fault probabilities. Specifically, based on the fault frequency features generated in step S810, a time series prediction model (such as exponential smoothing, an extension of the ARIMA model, etc.) is used to predict the next fault time for each faulty functional unit. Then, the difference between the predicted next fault time and the current time is calculated, and a first fault probability is generated based on this difference. The smaller the difference, the higher the first fault probability, indicating a greater likelihood that the faulty functional unit will fail in the near future.

[0053] Step S830 involves collecting real-time operational data from the tippler unhooking robot, analyzing the operational scenario and load characteristics, and determining the probability of failure for the multiple faulty functional units under the current operational scenario and load characteristics. The multiple first fault probabilities are then compensated to generate the multiple fault probabilities. Specifically, sensors and a monitoring system are used to collect real-time operational data from the tippler unhooking robot, including operational scenarios (such as working environment, temperature, humidity, etc.) and load characteristics (such as load size, load type, load distribution, etc.). By constructing a conditional probability model (a model describing the probability of an event occurring under certain conditions) and applying Bayesian networks (a model based on probability theory and graph theory used to represent dependencies between variables and for inference and prediction), this data is used to analyze the impact of the current operational scenario and load characteristics on the probability of failure for each faulty functional unit. Based on the analysis results, the first fault probabilities generated in step S820 are compensated to generate more accurate fault probabilities. This implementation method, by combining historical fault record datasets and real-time operational data, comprehensively analyzes and predicts the fault probabilities of faulty functional units, improving the accuracy of prediction.

[0054] In one possible implementation, step S830 further includes step S831, which analyzes the failure probability of the multiple faulty functional units under the current operating scenario and load characteristics, and analyzes the frequency distribution characteristics of the faulty operating scenarios and faulty load characteristics of the multiple faulty functional units based on the multiple historical fault record datasets. Specifically, data mining and statistical analysis methods are used to process the multiple historical fault record datasets. For each faulty functional unit, the operating scenario and load characteristic information in its historical fault records are extracted, and the frequency of these characteristics in different fault events is counted. For example, a frequency distribution table or frequency distribution graph can be constructed, where each feature combination corresponds to a frequency value, representing the number of times the feature combination occurs in the fault event.

[0055] Step S832: Based on the frequency distribution characteristics, calculate the proportion of the corresponding distribution frequency of the work scenario and load characteristics to the total fault frequency in the frequency distribution characteristics, and generate the failure probability of the multiple fault functional units under the current work scenario and load characteristics. Specifically, based on step S831, calculate the proportion of the frequency of each work scenario and load characteristic to the total fault frequency. This proportion is an estimate of the probability of the fault functional unit failing under the current work scenario and load characteristics. That is, for each fault functional unit, divide the frequency of each feature combination in its historical fault records by the total fault frequency (the total number of times a fault occurs) to obtain the failure probability corresponding to that feature combination. Then, for the work scenario and load characteristics in the current real-time work data, find the matching feature combination in the frequency distribution characteristics and obtain its corresponding failure probability. This probability value reflects the possibility of the fault functional unit failing under the current work scenario and load characteristics. This implementation method, by comprehensively considering the operational scenarios and load characteristics information in historical fault records, as well as this information in current real-time operational data, accurately estimates the probability of a faulty functional unit failing under the current operational scenarios and load characteristics, thereby improving the accuracy of fault analysis.

[0056] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A real-time data-driven fault analysis method for a tipper unhooking robot, characterized in that, include: Identify the multiple functional units of the tipper unhooking robot; Establish the normal collaborative relationships among the multiple functional units, as well as the set of abnormal collaborative impact relationships when any one or more units malfunction; The set of abnormal collaborative influence relationships is used to perform a discrimination analysis of fault features, and M functional unit combinations with discrimination that meet the preset discrimination index are established, where M is an integer greater than 1. A fault detection model is constructed based on the abnormal collaborative influence relationships of each functional unit in the combination of M functional units and the normal collaborative relationships. The system collects real-time operating data from the multiple functional units and inputs it into the fault detection model for fault analysis, then outputs a fault warning signal. The fault detection model is constructed using the abnormal collaborative influence relationships and normal collaborative relationships corresponding to each functional unit in the M functional unit combinations, including: For each abnormal collaborative influence relationship and the normal collaborative relationship corresponding to each functional unit in the M functional unit combination, a topology transformation is performed to generate each abnormal collaborative influence topology and the normal collaborative relationship topology. Calculate the topological difference vectors between each of the abnormal cooperative influence topologies and the normal cooperative relationship topologies; The fault detection model is trained using the aforementioned topological difference vectors; Specifically, the set of abnormal collaborative influence relationships is used to perform fault feature discrimination analysis, and M functional unit combinations with discrimination scores satisfying preset discrimination indices are established, including: Extract the first abnormal collaborative influence relationship and the second abnormal collaborative influence relationship from the set of abnormal collaborative influence relationships, as well as the corresponding first abnormal unit labeling information and second abnormal unit labeling information. The abnormal collaborative influence relationship refers to the collaborative relationship exhibited by two or more functional units under abnormal conditions. The first abnormal collaborative influence relationship and the second abnormal collaborative influence relationship refer to any two different abnormal collaborative influence relationships in the set of abnormal collaborative influence relationships. Perform a discrimination analysis on the collaboration relationship data between functional units in the first and second abnormal collaboration relationships to generate a first discrimination score; If the first discrimination index is less than the preset discrimination index, the marked abnormal units in the first abnormal unit marking information and the second abnormal unit marking information are merged into a functional unit combination and added to the M functional unit combinations. If the first discrimination index is greater than or equal to the preset discrimination index, the marked abnormal units in the first abnormal unit marking information and the second abnormal unit marking information are respectively combined as functional units and added to the M functional unit combinations. The determination of the preset discrimination index includes: Determine the preset network structure for constructing the fault detection model; Configure a binary classifier using the preset network structure, and construct several sets of collaborative relationship sample data corresponding to several data discrimination degrees; The binary classifier is tested using the aforementioned sets of collaborative relationship sample data, and the minimum discrimination score with a test classification accuracy greater than a preset threshold is obtained and set as the preset discrimination score index.

2. The real-time data-driven fault analysis method for a tipper unhooking robot as described in claim 1, characterized in that, After outputting the fault warning signal, it also includes: Determine whether the fault warning signal includes warnings for multiple faulty functional units; If so, connect to the fault log database to collect multiple historical fault record datasets from the multiple fault functional units; The failure probability at the current moment is analyzed using the multiple historical failure record datasets to generate multiple failure probabilities corresponding to the multiple failure functional units; Based on the multiple fault probabilities, they are arranged in descending order to generate a recommended maintenance priority sequence, which is sent to the control terminal of the tipper unhooking robot together with the fault warning signal.

3. The real-time data-driven fault analysis method for a tipper unhooking robot as described in claim 2, characterized in that, Using the multiple historical fault record datasets, perform fault probability analysis at the current moment to generate multiple fault probabilities corresponding to the multiple faulty functional units, including: Based on the multiple historical fault record datasets, the frequency of faults in the multiple fault functional units is analyzed to generate multiple fault frequency features. The difference between the predicted next fault time and the current time is used to generate multiple first fault probabilities based on the multiple fault frequency features; Real-time operation data of the tipper unhooking robot is collected, and the operation scenario and load characteristics are analyzed. The failure probability of the multiple fault functional units under the current operation scenario and load characteristics is analyzed, and the multiple first failure probabilities are compensated to generate the multiple failure probabilities.

4. The real-time data-driven fault analysis method for a tipper unhooking robot as described in claim 3, characterized in that, Analyze the probability of failure of the multiple fault-prone functional units under the current operating scenario and load characteristics, including: Based on the multiple historical fault record datasets, analyze the frequency distribution characteristics of the fault operation scenarios and fault load characteristics of the multiple fault functional units. Based on the frequency distribution characteristics, the proportion of the corresponding distribution frequency of the work scenario and load characteristics to the total fault frequency in the frequency distribution characteristics is calculated, and the failure probability of the multiple fault functional units under the current work scenario and load characteristics is generated.

5. The real-time data-driven fault analysis method for a tipper unhooking robot as described in claim 1, characterized in that, The multiple functional units include at least the robotic arm, gripping device, vision system, control system, and drive system of the tipper unhooking robot.

Images