Running state monitoring system and method for ultra-silence linear guide rail

Abstract

The invention provides a running state monitoring system and method for an ultra-silence linear guide rail. The method comprises the steps that graph structure data representing the incidence relation between running data of the ultra-silence linear guide rail is constructed through multi-source running data of the ultra-silence linear guide rail in the running process; constructing a guide rail health reference model including ultra-silence linear guide rail spatial-temporal feature extraction capability, and inputting the graph structure data into the guide rail health reference model for processing so as to extract state representation vectors corresponding to the operation data including the spatial-temporal dependency relationship; determining predicted health state information of the ultra-silence linear guide rail based on the structural characteristics and the state representation vector of the ultra-silence linear guide rail; and evaluating the running state of the ultra-silence linear guide rail according to the predicted health state information. By adopting the scheme of the invention, the operation state of the ultra-silence linear guide rail can be monitored based on the space-time dependency relationship among the operation data of the ultra-silence linear guide rail.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of running state monitoring, and more particularly to a running state monitoring system and method for a super-silent linear guide rail. BACKGROUND

[0002] Running state monitoring refers to a process of continuously tracking and evaluating real-time key parameters of equipment, systems or processes through sensors, data acquisition systems and analysis tools. Its core lies in timely capturing abnormal features, identifying performance degradation trends, and achieving early fault warning, thereby supporting predictive maintenance, improving operation reliability, safety and efficiency, and reducing unexpected downtime and maintenance costs.

[0003] As a precision transmission core component with low vibration and low noise characteristics, the super-silent linear guide rail is widely used in precision machinery, intelligent equipment and other types of equipment with high requirements for transmission accuracy and operation stability. Its running state directly determines the overall transmission performance and service life of the equipment, so accurate and comprehensive running state monitoring of the super-silent linear guide rail has become a core requirement for intelligent operation and maintenance of the equipment. However, existing guide rail running state monitoring methods mostly use single-dimensional data acquisition and analysis mode, do not model the internal correlation between vibration, noise, displacement and other multi-source running data, and most monitoring methods ignore the time-space dependent characteristics of guide rail running data evolution over time and spatial layout coupling. The health state determination link also does not fully consider the structural characteristics of the guide rail and the differentiated influence of each structural unit, but only relies on simple judgment based on single parameter threshold. Meanwhile, the running state evaluation lacks a multi-dimensional standardized index system tailored to the special characteristics of super-silent guide rails, resulting in problems such as incomplete feature extraction, large deviation between health determination and actual state, one-sided evaluation results and weak guidance for on-site operation and maintenance. Therefore, it is difficult to accurately capture early degradation signals and potential fault risks of the guide rail, and it is not suitable for the special operation requirements of super-silent linear guide rails and the actual needs of equipment intelligent operation and maintenance. Therefore, how to monitor the running state of the super-silent linear guide rail based on the space-time dependent relationship between the running data of the super-silent linear guide rail has become a problem in the industry. SUMMARY

[0004] The present application provides a running state monitoring system and method for a super-silent linear guide rail, which can monitor the running state of the super-silent linear guide rail based on the space-time dependent relationship between the running data of the super-silent linear guide rail.

[0005] In a first aspect, the present application provides a running state monitoring method for a super-silent linear guide rail, comprising the following steps: Collecting multi-source running data of the super-silent linear guide rail during operation; Based on the multi-source operating data, a graph structure data representing the correlation between the operating data of the ultra-quiet linear guide is constructed; A guide rail health benchmark model is constructed, which includes the ability to extract the spatiotemporal features of the ultra-quiet linear guide rail. The graph structure data is input into the guide rail health benchmark model for processing to extract the state representation vectors corresponding to each running data containing spatiotemporal dependencies. Based on the structural characteristics of the ultra-quiet linear guide and the state representation vector, the health status of the ultra-quiet linear guide is predicted to obtain the predicted health status information of the ultra-quiet linear guide. The operating status of the ultra-quiet linear guide is assessed based on the predicted health status information.

[0006] In some embodiments, the multi-source operating data includes high-precision micro-vibration, temperature, laser displacement, low-noise sound pickup physical sensor data, as well as operating parameters of the guide rail drive system such as operating speed, load, and operating time.

[0007] In some embodiments, constructing graph-structured data representing the correlation between the operating data of the ultra-quiet linear guide based on the multi-source operating data specifically includes: The multi-source operational data is preprocessed to obtain preprocessed multi-source operational data; The data association relationships between the various operational data are determined based on the preprocessed multi-source operational data. Using the collection points corresponding to each running data in the multi-source running data as nodes, and based on the data correlation between each running data, a graph structure data representing the correlation between the running data of the ultra-quiet linear guide is constructed.

[0008] In some embodiments, constructing a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail specifically includes: Obtain historical spatiotemporal characteristics of ultra-quiet linear guides throughout their entire lifecycle under rated load, different operating speeds, and different operating durations; The original historical spatiotemporal feature data is preprocessed to obtain preprocessed original historical spatiotemporal feature data. The preprocessed historical spatiotemporal feature raw data is divided into a training set, a validation set, and a test set to construct a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail.

[0009] In some embodiments, inputting the graph structure data into the guide rail health benchmark model for processing to extract the state representation vectors corresponding to each running data point containing spatiotemporal dependencies specifically includes: The graph structure data is preprocessed, and the preprocessed graph structure data is input into the guide rail health benchmark model for further processing to extract the time and space dependencies of each running data. Based on the temporal and spatial dependencies of each piece of running data, determine the state representation vector corresponding to each piece of running data that contains temporal and spatial dependencies.

[0010] In some embodiments, predicting the health status of the ultra-quiet linear guide based on its structural characteristics and the state representation vector, and obtaining the predicted health status information of the ultra-quiet linear guide, specifically includes: Obtain the structural characteristics of the ultra-quiet linear guide; Determine the influence characteristics of each structural unit on the overall health status of the ultra-quiet linear guide rail; The initial health status information of the ultra-quiet linear guide is determined based on the structural characteristics and the state representation vector. The initial health status information is adjusted by considering all the influencing features to obtain the predicted health status information of the ultra-quiet linear guide.

[0011] In some embodiments, assessing the operating status of the ultra-quiet linear guide using the predicted health status information specifically includes: Construct a multi-dimensional evaluation index system for the ultra-quiet linear guide; The predicted health status information is mapped to the multi-dimensional evaluation index system to evaluate the operating status of the ultra-quiet linear guide.

[0012] Secondly, this application provides an operating status monitoring system for ultra-quiet linear guides, comprising: The data acquisition module is used to collect multi-source operating data of the ultra-quiet linear guide during operation. The processing module is used to construct a graph structure data representing the correlation between the operating data of the ultra-quiet linear guide based on the multi-source operating data; The processing module is also used to construct a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail, and input the graph structure data into the guide rail health benchmark model for processing, so as to extract the state representation vector corresponding to each running data containing spatiotemporal dependencies. The processing module is also used to predict the health status of the ultra-quiet linear guide based on the structural characteristics of the ultra-quiet linear guide and the state representation vector, so as to obtain the predicted health status information of the ultra-quiet linear guide. An execution module is used to evaluate the operating status of the ultra-quiet linear guide based on the predicted health status information.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described method for monitoring the operating status of ultra-quiet linear guides.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for monitoring the operating status of an ultra-quiet linear guide.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The system and method for monitoring the operating status of ultra-quiet linear guides provided in this application collect multi-source operating data during the operation of the ultra-quiet linear guides, providing complete and fundamental raw data support for mining the spatiotemporal dependencies between multi-source data. Based on the multi-source operating data, a graph structure data is constructed, which can characterize the correlation between various operating data, providing a suitable structural carrier for extracting spatiotemporal features. A guideway health benchmark model with spatiotemporal feature extraction capabilities is constructed and input into the graph structure data to extract state representation vectors containing spatiotemporal dependencies, achieving accurate mining and quantitative characterization of the spatiotemporal dependencies of multi-source operating data. The health status is predicted by combining the guideway structural characteristics and the state representation vectors, ensuring that the health status prediction aligns with the actual structural attributes of the guideway. The accuracy and relevance of the prediction results are improved by relying on the mined spatiotemporal dependencies. The predicted health status information is used to assess the guideway operating status, and based on the aforementioned mined spatiotemporal dependencies, an accurate determination of the ultra-quiet linear guideway operating status is achieved, realizing the core objective of monitoring guideway operating status based on spatiotemporal dependencies. By adopting the solution of this application, the operating status of the ultra-quiet linear guide can be monitored based on the spatiotemporal dependency between various operating data of the ultra-quiet linear guide. Attached Figure Description

[0016] Figure 1 This is an exemplary flowchart of a method for monitoring the operating status of an ultra-quiet linear guide, according to some embodiments of this application. Figure 2 This is an exemplary flowchart illustrating the determination of graph structure data according to some embodiments of this application; Figure 3 This is an exemplary flowchart illustrating the determination of predicted health status information according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of an operating status monitoring system for ultra-quiet linear guides, according to some embodiments of this application; Figure 5This is a schematic diagram of the structure of a computer device for implementing a method for monitoring the operating status of an ultra-quiet linear guide, according to some embodiments of this application. Detailed Implementation

[0017] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] refer to Figure 1 The figure is an exemplary flowchart of a method for monitoring the operating status of an ultra-quiet linear guide according to some embodiments of this application. The method for monitoring the operating status of an ultra-quiet linear guide mainly includes the following steps: In step 101, multi-source operating data of the ultra-quiet linear guide rail during operation are collected.

[0019] It should be noted that the multi-source operating data in this application is a collection of multi-type, time-series, and multi-point operating data synchronously collected from key components and preset collection points of different monitoring dimensions of the ultra-quiet linear guide, reflecting the real-time physical state, operating condition characteristics, and dynamic correlation between various monitoring parameters of the key components of the guide body, slider, and raceway during the operation of the ultra-quiet linear guide. It can accurately capture the micro-scale state changes of the guide under ultra-quiet conditions. The multi-source operating data includes high-precision physical sensor data such as micro-vibration, temperature, laser displacement, and low-noise sound pickup, as well as operating condition parameters of the guide drive system such as operating speed, load, and running time.

[0020] In step 102, a graph structure data representing the correlation between the operating data of the ultra-quiet linear guide is constructed based on the multi-source operating data.

[0021] In some embodiments, reference Figure 2 The diagram is an exemplary flowchart for determining graph structure data in some embodiments of this application. In this embodiment, the graph structure data representing the correlation between the operating data of the ultra-quiet linear guide can be constructed based on the multi-source operating data using the following steps: In step 1021, the multi-source operating data is preprocessed to obtain preprocessed multi-source operating data; In step 1022, the data association relationship between each piece of operational data is determined based on the preprocessed multi-source operational data; In step 1023, the corresponding acquisition points of each running data in the multi-source running data are used as nodes, and a graph structure data representing the relationship between the running data of the ultra-quiet linear guide is constructed based on the data association relationship between each running data.

[0022] In specific implementation, the multi-source operational data is preprocessed to obtain the preprocessed multi-source operational data as follows: First, the multi-source operational data is processed in its entirety using a data preprocessing method. Outliers in the original data are identified and removed by setting a significance level of 0.05 using the Grubbs test. For missing data locations, linear interpolation is used to fill in missing values ​​based on the changing trends of adjacent time-series data. For vibration and noise signal data, a db4 wavelet basis is selected for 3-level wavelet decomposition, and high-frequency coefficients are processed using soft thresholding to achieve signal denoising. Then, the min-max normalization method is used to map the operational data of each dimension to the [0,1] interval to eliminate dimensional differences. Finally, the preprocessed multi-source operational data that is standardized, time-seriesd, and free of redundant interference is obtained.

[0023] In addition, in specific implementation, the data association relationship between each set of operational data is determined based on the preprocessed multi-source operational data as follows: The Pearson correlation coefficient method is used to calculate the ratio of the covariance to the standard deviation of any two sets of operational data sequences to obtain the correlation degree value in the [-1,1] interval, which represents the degree of linear association. At the same time, the mutual information method is used to calculate the difference between the joint entropy and the marginal entropy of any two sets of operational data sequences to obtain the non-negative correlation degree value, which represents the degree of nonlinear information interaction. Then, a known correlation degree judgment threshold is set, and the correlation relationship with an absolute value of Pearson correlation coefficient ≥ 0.3 and a mutual information value ≥ 0.1 is judged as a valid data association relationship. The quantitative correlation strength value corresponding to each valid correlation relationship is retained to obtain the data association relationship between each set of operational data.

[0024] In addition, in specific implementation, the graph structure data representing the correlation between the ultra-quiet linear guide rail operation data is constructed based on the data correlation between each operation data, using the corresponding acquisition point of each operation data in the multi-source operation data as the node. Specifically, the graph structure data is constructed by using the corresponding acquisition point of each operation data in the multi-source operation data as the node, assigning the pre-processed full-time sequence operation data of the corresponding acquisition point as the node feature, establishing connection edges between corresponding nodes based on the aforementioned effective data correlation, using the absolute value of the Pearson correlation coefficient or the mutual information value corresponding to each effective correlation as the weight of the corresponding edge to quantify the correlation strength between nodes, then using the adjacency matrix to matrix-represent the connection relationship and corresponding weight between nodes, and using the node feature matrix to matrix-represent the temporal feature information of each node. Finally, a graph structure data that can reflect the spatial correlation and quantitative correlation strength between the ultra-quiet linear guide rail operation data is constructed by the node feature matrix and the adjacency matrix.

[0025] It should be noted that the data association relationships in this application represent the logical connections, quantitative correspondences, and inherent relationships of mutual constraints between different dimensions of data in the garden pest inspection area. They reflect the mutual influence and synergistic relationships of various types of data in the inspection path planning scenario, and clearly define the association logic and quantitative matching relationships between different data. The graph structure data reflects the topological distribution relationship of inspection-related core objects and the structured characteristics of the association attributes between objects, and intuitively presents the path connectivity topology between clustered areas and the carrier distribution state of various association parameters.

[0026] In step 103, a guide rail health benchmark model is constructed, which includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail. The graph structure data is input into the guide rail health benchmark model for processing to extract the state representation vectors corresponding to each running data containing spatiotemporal dependencies.

[0027] In some embodiments, constructing a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail can be achieved by the following steps: Obtain historical spatiotemporal characteristics of ultra-quiet linear guides throughout their entire lifecycle under rated load, different operating speeds, and different operating durations; The original historical spatiotemporal feature data is preprocessed to obtain preprocessed original historical spatiotemporal feature data. The preprocessed historical spatiotemporal feature raw data is divided into a training set, a validation set, and a test set to construct a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail.

[0028] It should be noted that the historical spatiotemporal feature raw data in this application represents the original monitoring data of spatial dimension monitoring characteristics and temporal dimension change characteristics of the ultra-quiet linear guide during its entire life cycle operation under rated load, different operating speeds, and different operating durations. It reflects the real operating status of the ultra-quiet linear guide under different operating conditions, the dynamic evolution of physical parameters at each monitoring point with operating time, and the spatiotemporal operating law of the guide from healthy operation to slight degradation and early failure throughout its entire life cycle. The historical spatiotemporal feature raw data includes the original values ​​of spatial dimension physical parameters of each preset monitoring point on the guide slider and guide surface under rated load, different operating speeds, and different operating durations throughout its entire life cycle, namely vibration amplitude, noise decibel value, displacement deviation, friction force, and surface temperature value, as well as the time-series raw data of the continuous change of the above parameters at each point with operating time. It also includes the original labeling of the guide operating condition and the original labeling of the guide health status corresponding to the data acquisition time.

[0029] In specific implementation, the original historical spatiotemporal feature data is preprocessed to obtain the preprocessed original historical spatiotemporal feature data as follows: Data cleaning methods are used to remove extreme outliers using the 3σ principle, and linear interpolation is used to supplement a small amount of missing data, completing the initial data purification; for time-series data, downsampling is used to uniformly convert high-frequency time-series data to a fixed sampling frequency, eliminating the time-series disorder caused by different acquisition frequencies, and time-series synchronization processing is performed to ensure that the time axis of the time-series data at each monitoring point is consistent; for spatial data, coordinate registration is used to calibrate the spatial coordinates of each monitoring point to ensure accurate point location correspondence, and then minimum-maximum normalization is used to uniformly map spatial feature data of different dimensions such as vibration, noise, and displacement to the 0-1 interval, completing the data dimension unification and standardization, and finally obtaining the preprocessed original historical spatiotemporal feature data.

[0030] In addition, in specific implementation, the preprocessed historical spatiotemporal feature raw data is divided into training set, validation set, and test set to construct a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail. Specifically, the preprocessed historical spatiotemporal feature raw data is randomly divided into training set, validation set, and test set in a 7:2:1 ratio. A guide rail health benchmark model is constructed based on the existing deep learning infrastructure. The core of the model embeds a spatiotemporal feature extraction module. This module uses a convolutional neural network as the spatial feature extraction unit. It inputs standardized spatial feature data and extracts the spatial correlation features of each monitoring point through convolution and pooling operations. It uses a long short-term memory network as the temporal feature extraction unit. It inputs synchronized temporal feature data and extracts the temporal trend features of parameters changing over time. The outputs of the two are then spliced ​​and fused. Subsequently, high-dimensional spatiotemporal fusion features of the guide rail are obtained. The main body of the model uses fully connected layers to construct a health benchmark fitting unit. Taking the spatiotemporal fusion features as input, it outputs a quantitative benchmark value of the guide rail health status. During training, gradient descent is used to optimize the model's cross-entropy loss function. The training set drives the model to learn the intrinsic relationship between spatiotemporal features and guide rail health status. The model's spatiotemporal feature extraction accuracy and health benchmark fitting effect are monitored in real time through the validation set. The model network parameters are adjusted, and the generalization ability of the model is verified using the test set. When the model's spatiotemporal feature extraction accuracy on the test set reaches more than 95%, and the health benchmark fitting error converges to the preset threshold and stabilizes, training is stopped. Finally, a guide rail health benchmark model with the ability to extract spatiotemporal features of ultra-quiet linear guide rails is constructed. Other methods can be used in other embodiments, which are not limited here.

[0031] It should be noted that the guide rail health benchmark model in this application represents a condition-adaptive quantitative judgment model for the health status of ultra-quiet linear guides, capable of extracting spatiotemporal features. It serves as a digital benchmark carrier for the health status of ultra-quiet linear guides, reflecting the inherent quantitative correlation between spatiotemporal characteristic parameters and guide rail health status throughout the entire life cycle of ultra-quiet linear guides under specific operating conditions such as rated load, different operating speeds, and different operating durations. It also reflects the spatiotemporal characteristic benchmark thresholds and parameter change trends under different states such as healthy operation, minor degradation, and early failures of this type of guide rail, intuitively demonstrating the correspondence between the dynamic evolution of guide rail health status and spatiotemporal features. This model can be used for feature extraction and quantitative analysis of real-time acquired spatiotemporal characteristic data of ultra-quiet linear guides.

[0032] In some embodiments, the graph structure data is input into the guide rail health benchmark model for processing to extract the state representation vectors corresponding to each running data containing spatiotemporal dependencies. This can be achieved through the following steps: The graph structure data is preprocessed, and the preprocessed graph structure data is input into the guide rail health benchmark model for further processing to extract the time and space dependencies of each running data. Based on the temporal and spatial dependencies of each piece of running data, determine the state representation vector corresponding to each piece of running data that contains temporal and spatial dependencies.

[0033] In specific implementation, the graph structure data is preprocessed, and the preprocessed graph structure data is input into the guide rail health benchmark model for further processing. The extraction of the time and spatial dependencies of each operational data point involves: systematically preprocessing the graph structure data of the ultra-quiet linear guide rail. This graph structure data uses each operational monitoring point of the guide rail as nodes and the spatiotemporal correlation between points as edges. Node attributes include all-dimensional operational data such as vibration, noise, displacement, and friction, as well as time-series acquisition labels. Edge attributes include the spatial position dependency between points and the time-series operational parameters. For order correlation, the preprocessing first employs a well-known outlier removal method, filtering extreme abnormal operational data in node attributes using the 3σ principle. Invalid nodes without actual monitoring data and redundant edges without spatiotemporal correlation are removed from the graph structure. Then, the minimum-maximum normalization method is used to unify the dimensions of various node operational data, mapping the data to the 0-1 interval. The adjacency matrix representing the correlation between points is standardized using Laplace normalization to eliminate matrix scale differences. Finally, the processed node feature matrix and normalized adjacency matrix are uniformly converted into a guide rail health benchmark model. The model uses a recognizable three-dimensional tensor format to obtain preprocessed graph structure data with regular structure and unified features. The preprocessed graph structure data is then input into a converged guide rail health benchmark model. The model's built-in spatiotemporal feature extraction unit sequentially extracts the spatial and temporal dependencies of each running data point. For spatial dependency extraction, a basic convolutional layer of a graph convolutional neural network is used, taking the normalized adjacency matrix and node feature matrix as input. A preset multidimensional convolutional kernel performs neighborhood aggregation operations on the node features to capture the coupling features of running parameters caused by spatial location associations between different monitoring points, outputting the spatial dependency feature vector of each running data point. This spatial dependency feature vector represents the spatial dependency. For temporal dependency extraction, a time-series convolutional layer is used, taking the time-series sequence of running data sorted by acquisition time as input. A sliding operation of a one-dimensional convolutional kernel extracts the trend change features and temporal correlation features of the running data in the time dimension, outputting the time dependency feature vector of each running data point. This time dependency feature vector represents the time dependency. Other implementation methods can be used in other embodiments, which are not limited here.

[0034] In addition, in specific implementation, the determination of the state representation vector corresponding to each running data containing spatiotemporal dependencies based on the temporal and spatial dependencies of each running data is as follows: The temporal dependency feature vector corresponding to the temporal dependency and the spatial dependency feature vector corresponding to the spatial dependency of each running data are subjected to feature fusion processing. The two types of feature vectors of running data along the same path are sequentially concatenated according to their dimensions using a feature concatenation method to obtain a high-dimensional spatiotemporal fusion feature vector. Then, principal component analysis is used to perform dimensionality reduction optimization on the high-dimensional feature vector, eliminating redundant features and retaining the core principal components representing spatiotemporal dependencies. The dimensionality-reduced feature vector is then normalized using the Z-score standardization method to calibrate the mean and variance, resulting in a feature vector with fixed dimensions and consistent feature recognition. This feature vector is then used to determine the state representation vector corresponding to each running data containing spatiotemporal dependencies, ensuring that each running data has a unique and complete state representation vector that fully represents its spatiotemporal dependency characteristics. Other implementation methods can also be used in other embodiments, which are not limited here.

[0035] It should be noted that the time dependency relationship in this application represents the inherent correlation law of the time-series coupling, trend evolution, and temporal coupling of the various operating data of the ultra-quiet linear guide. It reflects the temporal characteristics of the dynamic changes of each operating data with the running time and operating stage of the guide and the mutual constraints between data in the time dimension. It can be used to extract the temporal change trend of the operating data and capture the dynamic evolution law of the guide's operating state in the time dimension. The spatial dependency relationship represents the inherent correlation law of the coupling of operating data and parameter linkage between different monitoring points of the ultra-quiet linear guide due to the spatial layout and the overall structural relationship of the guide. It reflects the spatial dimension mutual influence relationship of the operating data of each point due to the structural characteristics of the guide. It can be used to capture the coordinated change characteristics of operating parameters between different points and reflect the spatial distribution law of the overall operating state of the guide. The state representation vector reflects the comprehensive characteristics of each operating data that have both temporal evolution and spatial linkage and the corresponding local or overall operating state of the guide. It can be used to accurately characterize the spatiotemporal essential characteristics of the guide's operating data and provide standardized and quantifiable feature inputs for subsequent work such as guide health status determination and early fault warning.

[0036] In step 104, the health status of the ultra-quiet linear guide is predicted based on the structural characteristics of the ultra-quiet linear guide and the state characterization vector, thereby obtaining the predicted health status information of the ultra-quiet linear guide.

[0037] In some embodiments, reference Figure 3The figure is an exemplary flowchart for determining predicted health status information in some embodiments of this application. In this embodiment, the health status of the ultra-quiet linear guide is predicted based on the structural characteristics of the ultra-quiet linear guide and the state representation vector. The predicted health status information of the ultra-quiet linear guide can be obtained by the following steps: In step 1041, the structural characteristics of the ultra-quiet linear guide are obtained; In step 1042, the influence characteristics of each structural unit on the ultra-quiet linear guide rail on the overall health status of the guide rail are determined; In step 1043, the initial health status information of the ultra-quiet linear guide is determined based on the structural characteristics and the state characterization vector. In step 1044, the initial health status information is adjusted using all the influencing features to obtain the predicted health status information of the ultra-quiet linear guide.

[0038] In specific implementation, obtaining the structural characteristics of the ultra-quiet linear guide rail involves: retrieving the design drawings, factory calibration parameters, and structural processing and testing reports of the ultra-quiet linear guide rail, and combining them with measured structural data under actual operating scenarios, to obtain its complete structural characteristics. Specifically, this includes the overall geometric parameters of the ultra-quiet linear guide rail, the mechanical transmission characteristics of each component, and the core technical points of the quiet design. Simultaneously, the guide rail is disassembled into independent structural units such as the slider, the core contact area of ​​the guide rail surface, the raceway assembly, the lubrication system, the sealing end cap, and the mounting base connection area. The spatial layout, functional attributes, and correlation with the quiet operation and overall transmission performance of each structural unit, as well as the assembly and coordination characteristics and operational linkage patterns between units, are analyzed one by one. Each structural unit is classified and labeled according to its core function, important auxiliary function, and basic support function, forming the structural characteristics of the ultra-quiet linear guide rail that includes unit attributes, structural parameters, and correlations. Other methods can be used in other embodiments, which are not limited here.

[0039] Furthermore, in specific implementation, determining the influence characteristics of each structural unit on the overall health status of the ultra-quiet linear guide rail is as follows: Using the analytic hierarchy process (AHP) combined with historical operating data of the ultra-quiet linear guide rail, the influence characteristics of each structural unit on the overall health status of the guide rail are determined. First, an AHP structural model is constructed, with the overall health status of the guide rail set as the target layer. The functional importance of the structural units, the historical probability of degradation leading to overall failure, and the degree of coupling between the operating states of each unit are set as the criterion layer. Each structural unit is set as the scheme layer. Judgment matrices for the criterion layer and the scheme layer are constructed. The judgment matrices are normalized and subjected to consistency checks. When the consistency ratio is less than 0.1, the quantitative influence weight of each structural unit is determined. After normalization, the impact dimensions, fault correlation characteristics, and state coupling characteristics of each structural unit are extracted by statistically analyzing historical degradation data. The impact dimensions include the types and quantitative proportions of the impact of parameters such as vibration, noise, displacement, friction, and temperature on the overall health. The fault correlation characteristics include the degree and probability of the correlation between the degradation of a single unit and the failure of other units or the overall guide rail. The state coupling characteristics include the degree of mutual constraint of the operating states of each structural unit. The coupling coefficient is calculated by cross-correlation analysis. The above impact dimensions, fault correlation characteristics, and state coupling characteristics are used as the standardized and quantifiable impact characteristics of the corresponding structural units. Other methods can be used in other embodiments, which are not limited here.

[0040] In addition, in specific implementation, determining the initial health status information of the ultra-quiet linear guide rail based on the structural characteristics and the state representation vectors is as follows: Based on the spatial affiliation and functional relationships of each structural unit in the guide rail's structural characteristics, the state representation vectors corresponding to each operating data point are mapped to the corresponding structural units according to the spatial coordinates and functional attributes of the monitoring points, forming a dedicated state representation vector set for each structural unit. The vector sets of each unit are preprocessed, and low-discrimination redundant feature vectors are removed through variance analysis. Principal component analysis is used to optimize the dimensionality of the remaining vectors, extracting the principal component feature vectors that can represent the core operating status of the unit. These principal component feature vectors are then input into the trained and converged guide rail health benchmark model and compared with the health benchmark feature library of each structural unit built into the model. The health benchmark feature library is divided into four categories: healthy, minor degradation, early failure, and failure. The health level is preset with corresponding baseline feature vectors and similarity judgment thresholds for each level. A dual matching method using cosine similarity and Euclidean distance is employed to obtain the matching degree between the vector to be judged and the baseline vectors of each level. The initial health level of each structural unit is determined based on the maximum matching degree, and each unit is assigned an initial health score according to a quantification standard of 0-100. Then, based on the functional importance weights of each structural unit in the guide rail's structural characteristics, a weighted summation method is used to aggregate the initial health scores of all structural units, obtaining the initial health score of the guide rail as a whole. The initial health level of the guide rail is determined by combining the score, and the initial health anomalies and anomaly parameter types of each structural unit are marked, forming the initial health status information of the ultra-quiet linear guide rail, which includes unit-level and overall health information. Other methods can be used in other embodiments, which are not limited here.

[0041] In addition, in specific implementation, adjusting the initial health status information based on all influencing features to obtain the predicted health status information of the ultra-quiet linear guide is specifically as follows: Using the influencing features of each structural unit as the basis for correction, the initial health status information is adjusted in a multi-dimensional and refined manner. First, based on the influence weights in the influencing features, the initial health scores of each structural unit are recalibrated and corrected to strengthen the influence ratio of the core unit on the overall health status. Then, combined with fault correlation characteristics, for structural units with minor degradation or early faults, the health scores of other structural units with the risk of linked faults are adjusted downward according to their fault correlation degree to avoid the judgment bias of single unit degradation. Subsequently, based on the quantized coupling coefficient in the state coupling feature of the influencing features, the operating state of structural units with high coupling degree is adjusted in a linked manner so that the unit health scores can reflect the actual operating state of mutual constraints. In conjunction with the silent design structure characteristics of the ultra-quiet linear guide, specific modifications are made to core units affecting the silent performance of the guide, such as the sealed end cap and the low-friction coating area on the guide surface. The scoring deviations corresponding to abnormal parameters of these units are amplified to ensure that the health status related to silent performance is accurately reflected. After all modifications are completed, the weighted summation method is used again to aggregate and calculate the health scores of each modified structural unit to obtain the overall health score and health level of the guide. At the same time, combined with the modified health scores of each structural unit and the degradation correlation rules in the influencing characteristics, the structural parts of the guide that may deteriorate, the degradation development rate and the predicted trend are analyzed. The health status, the overall health level and quantitative score, potential fault risk points, degradation trend prediction, and silent performance-related health risks are integrated as the predicted health status information of the ultra-quiet linear guide. Other methods can also be used in other embodiments, which are not limited here.

[0042] It should be noted that the structural characteristics in this application represent all inherent structural attributes and laws of the ultra-quiet linear guide, reflecting the inherent structural attributes of the guide, the functional positioning and spatial layout of each structural unit; the influence characteristics represent various quantitative attributes and correlation laws of each structural unit of the ultra-quiet linear guide affecting its overall health status, reflecting the degree and type of influence of each structural unit on the overall health of the guide; the initial health status information reflects the preliminary health status of each structural unit in the ultra-quiet linear guide during operation, which can be used as the basis for predicting the health status of the guide and clarifying the preliminary judgment direction of the health status of the guide; the predicted health status information reflects the true health status of the ultra-quiet linear guide after comprehensively considering its own structural characteristics and the mutual influence between each structural unit, which can be used to provide a comprehensive and quantitative judgment basis for the intelligent operation and maintenance of the ultra-quiet linear guide, guide early fault warning, accurate determination of operation and maintenance timing, formulation of personalized operation and maintenance strategies, and targeted detection and preventive maintenance of potentially deteriorated parts.

[0043] In step 105, the operating status of the ultra-quiet linear guide is evaluated using the predicted health status information.

[0044] In some embodiments, assessing the operating status of the ultra-quiet linear guide using the predicted health status information can be achieved through the following steps: Construct a multi-dimensional evaluation index system for the ultra-quiet linear guide; The predicted health status information is mapped to the multi-dimensional evaluation index system to evaluate the operating status of the ultra-quiet linear guide.

[0045] In specific implementation, the multi-dimensional evaluation index system for the ultra-quiet linear guide is constructed as follows: Combining the design characteristics of the ultra-quiet linear guide's quiet operation and high-precision transmission, and based on its factory rated operating parameters, industry operation and maintenance technical standards, and full life cycle operation requirements, a multi-dimensional evaluation index system is constructed using the analytic hierarchy process (AHP). First, the guide rail's operating status evaluation is divided into four primary evaluation dimensions: core operating performance, structural unit operating status, potential fault risks, and deterioration trends. Then, each primary dimension is further decomposed into indicators. Under the core operating performance dimension, secondary quantitative indicators directly related to quietness and transmission include vibration amplitude, noise decibels, displacement deviation, and friction magnitude. Under the structural unit operating status dimension, secondary quantitative indicators include the health compliance rate of core / important / auxiliary structural units and the proportion of abnormal points. Under the fault risk dimension, secondary quantitative indicators such as the number of risk points, risk level, and fault coupling probability are set. Under the deterioration development trend dimension, secondary quantitative indicators such as deterioration rate and linkage of deteriorated parts are set. Then, a judgment matrix for each level of indicators is constructed. The matrix is ​​assigned values ​​by expert scoring and consistency is checked. Under the premise that the consistency ratio is less than 0.1, the quantitative weight of each primary and secondary indicator is determined. The core operating performance dimension is given the highest weight. Then, according to the guide rail factory calibration parameters and industry standards, three-level quantitative judgment thresholds of rated compliance, early warning, and fault are set for each secondary indicator. At the same time, quantitative conversion rules for qualitative indicators are formulated to form a multi-dimensional evaluation indicator system with clear indicator hierarchy, reasonable weight allocation, and clear judgment standards. Other methods can be used in other embodiments, which are not limited here.

[0046] It should be noted that the multi-dimensional evaluation index system in this application reflects the evaluation orientation that conforms to the exclusive operating characteristics of ultra-quiet linear guides and industry standards, ensuring that the evaluation dimensions cover the core and key aspects of guide rail operation; this system can be used to provide a unified and standardized evaluation basis and analysis framework for the evaluation of the operating status of ultra-quiet linear guides, ensuring the standardization of the guide rail operating status evaluation process and the objectivity, accuracy and traceability of the evaluation results.

[0047] In addition, in specific implementation, mapping the predicted health status information with the multi-dimensional evaluation index system to evaluate the operating status of the ultra-quiet linear guide rail is specifically as follows: According to the preset mapping rule of "first-level evaluation dimension - second-level quantitative index - data attribute," each data point in the predicted health status information is matched one by one to the corresponding second-level index in the multi-dimensional evaluation index system. Qualitative information such as health level and risk level is converted into standardized quantitative scores of 0-100 points according to the quantitative conversion rules. Quantitative data such as health score and deterioration rate are directly substituted into the corresponding index for threshold comparison. The threshold comparison method is used to determine the rated compliance, warning, or fault range of each second-level index. Then, according to the index weights determined by the analytic hierarchy process, the weighted comprehensive scoring method is used to perform layer-by-layer weighted aggregation calculation of the quantitative scores of each second-level index, thereby obtaining the comprehensive scores of each first-level dimension. The quantitative comprehensive score of the overall operating status of the guide rail, combined with the interval judgment results of each indicator and the comprehensive score, analyzes whether the core operating performance meets the rated operating conditions, whether the structural unit has an unstable operating state, whether potential fault risks affect short-term operation, and whether the deterioration trend conforms to the normal evolution law of the whole life cycle. At the same time, it assesses the linkage and coupling effect between deteriorated parts and fault risk points. Finally, based on the preset four-level operating status classification standard of excellent (90-100 points), good (70-89 points), acceptable (50-69 points), and abnormal (<50 points), the overall operating status level of the ultra-quiet linear guide rail is calibrated, and the core compliance items, abnormal warning items and high-risk impact points in the evaluation process are identified, thus completing a comprehensive quantitative evaluation of the operating status of the ultra-quiet linear guide rail. Other methods can also be used in other embodiments, which are not limited here.

[0048] In another aspect, in some embodiments, this application provides an operating status monitoring system for ultra-quiet linear guides, referring to... Figure 4 The figure is a schematic diagram of the structure of a monitoring system for the operation status of an ultra-quiet linear guide rail according to some embodiments of this application. The monitoring system 400 for the operation status of an ultra-quiet linear guide rail includes: a data acquisition module 401, a processing module 402, and an execution module 403, which are described below: Acquisition module 401, in this application, is mainly used to acquire multi-source operating data of the ultra-quiet linear guide rail during operation; Processing module 402, in this application, is used to construct graph structure data representing the correlation between the operation data of the ultra-quiet linear guide based on the multi-source operation data; It should be noted that the processing module 402 in this application is also used to construct a guide rail health benchmark model including the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail, and input the graph structure data into the guide rail health benchmark model for processing, so as to extract the state representation vector corresponding to each running data containing spatiotemporal dependencies. In addition, it should be noted that the processing module 402 in this application is also used to predict the health status of the ultra-quiet linear guide based on the structural characteristics of the ultra-quiet linear guide and the state representation vector, so as to obtain the predicted health status information of the ultra-quiet linear guide. The execution module 403 in this application is mainly used to evaluate the operating status of the ultra-quiet linear guide based on the predicted health status information.

[0049] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described method for monitoring the operating status of ultra-quiet linear guides.

[0050] In some embodiments, reference Figure 5 This figure is a schematic diagram of the structure of a computer device implementing a method for monitoring the operating status of an ultra-quiet linear guide, according to some embodiments of this application. The method for monitoring the operating status of an ultra-quiet linear guide in the above embodiments can... Figure 5 The computer device shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0051] Processor 501 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0052] The communication bus 502 can be used to transmit information between the aforementioned components.

[0053] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 503 may exist independently and be connected to processor 501 via communication bus 502. Memory 503 may also be integrated with processor 501.

[0054] The memory 503 stores program code for executing the scheme of this application, and its execution is controlled by the processor 501. The processor 501 executes the program code stored in the memory 503. The program code may include one or more software modules. The method used in the above embodiments can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0055] Communication interface 504 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0056] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0057] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0058] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for monitoring the operating status of an ultra-quiet linear guide.

[0059] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0060] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for monitoring the operating status of ultra-quiet linear guides, characterized in that, Includes the following steps: Collect multi-source operating data of the ultra-quiet linear guide during operation; Based on the multi-source operating data, a graph structure data representing the correlation between the operating data of the ultra-quiet linear guide is constructed; A guide rail health benchmark model is constructed, which includes the ability to extract the spatiotemporal features of the ultra-quiet linear guide rail. The graph structure data is input into the guide rail health benchmark model for processing to extract the state representation vectors corresponding to each running data containing spatiotemporal dependencies. Based on the structural characteristics of the ultra-quiet linear guide and the state representation vector, the health status of the ultra-quiet linear guide is predicted to obtain the predicted health status information of the ultra-quiet linear guide. The operating status of the ultra-quiet linear guide is assessed based on the predicted health status information.

2. The method as described in claim 1, characterized in that, The multi-source operating data includes high-precision micro-vibration, temperature, laser displacement, and low-noise sound pickup physical sensor data, as well as operating parameters of the guide rail drive system such as operating speed, load, and operating time.

3. The method as described in claim 1, characterized in that, The graph-structured data representing the correlation between the operating data of the ultra-quiet linear guide rail, constructed based on the multi-source operating data, specifically includes: The multi-source operational data is preprocessed to obtain preprocessed multi-source operational data; The data association relationships between the various operational data are determined based on the preprocessed multi-source operational data. Using the collection points corresponding to each running data in the multi-source running data as nodes, and based on the data correlation between each running data, a graph structure data representing the correlation between the running data of the ultra-quiet linear guide is constructed.

4. The method as described in claim 1, characterized in that, The construction of a guide rail health benchmark model, including the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail, specifically includes: Obtain historical spatiotemporal characteristics of ultra-quiet linear guides throughout their entire lifecycle under rated load, different operating speeds, and different operating durations; The original historical spatiotemporal feature data is preprocessed to obtain preprocessed original historical spatiotemporal feature data. The preprocessed historical spatiotemporal feature raw data is divided into a training set, a validation set, and a test set to construct a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail.

5. The method as described in claim 1, characterized in that, The graph structure data is input into the guide rail health benchmark model for processing to extract the state representation vectors corresponding to each operational data point containing spatiotemporal dependencies. Specifically, this includes: The graph structure data is preprocessed, and the preprocessed graph structure data is input into the guide rail health benchmark model for further processing to extract the time and space dependencies of each running data. Based on the temporal and spatial dependencies of each piece of running data, determine the state representation vector corresponding to each piece of running data that contains temporal and spatial dependencies.

6. The method as described in claim 1, characterized in that, Based on the structural characteristics of the ultra-quiet linear guide and the state representation vector, the health status of the ultra-quiet linear guide is predicted, and the predicted health status information of the ultra-quiet linear guide specifically includes: Obtain the structural characteristics of the ultra-quiet linear guide; Determine the influence characteristics of each structural unit on the overall health status of the ultra-quiet linear guide rail; The initial health status information of the ultra-quiet linear guide is determined based on the structural characteristics and the state representation vector. The initial health status information is adjusted by considering all the influencing features to obtain the predicted health status information of the ultra-quiet linear guide.

7. The method as described in claim 1, characterized in that, The assessment of the operating status of the ultra-quiet linear guide based on the predicted health status information specifically includes: Construct a multi-dimensional evaluation index system for the ultra-quiet linear guide; The predicted health status information is mapped to the multi-dimensional evaluation index system to evaluate the operating status of the ultra-quiet linear guide.

8. A system for monitoring the operating status of ultra-quiet linear guides, characterized in that, include: The data acquisition module is used to collect multi-source operating data of the ultra-quiet linear guide during operation. The processing module is used to construct a graph structure data representing the correlation between the operating data of the ultra-quiet linear guide based on the multi-source operating data; The processing module is also used to construct a guide rail health benchmark model that includes the spatiotemporal feature extraction capability of the ultra-quiet linear guide rail, and input the graph structure data into the guide rail health benchmark model for processing, so as to extract the state representation vector corresponding to each running data containing spatiotemporal dependencies. The processing module is also used to predict the health status of the ultra-quiet linear guide based on the structural characteristics of the ultra-quiet linear guide and the state representation vector, so as to obtain the predicted health status information of the ultra-quiet linear guide. An execution module is used to evaluate the operating status of the ultra-quiet linear guide based on the predicted health status information.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the method for monitoring the operating status of an ultra-quiet linear guide as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for monitoring the operating status of ultra-quiet linear guides as described in any one of claims 1 to 7.

Images