Factory production intelligent supervision management method and system thereof

Abstract

The application discloses a kind of factory production intelligent supervision management method and system, it is related to production management technical field, including, based on operation behavior Construction time difference chain, extract fluctuation characteristic element by sliding window;Wave spectrum is obtained to rhythm fluctuation by the frequency domain transformation to fluctuation characteristic element;Dynamic time warping distance of rhythm fluctuation spectrum and historical rhythm reference spectrum is calculated, and spectrum deviation descriptor is obtained;Spectrum deviation descriptor is nested mapping in standard semantic process flow chart, and rhythm process nested chart is formed;Node level index semantic vector is extracted from rhythm process nested chart based on graph embedding algorithm, and multidimensional semantic alignment detection is executed, and alignment residual error measure is output;Alignment residual error measure is aggregated, and structured misplacement score list is generated, and abnormal behavior scene graph is constructed based on score threshold value.The application generates rhythm fluctuation spectrum and carries out semantic nested mapping, realizes the identification of exception and semantic misplacement in production supervision process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of production management, and in particular to a factory production intelligent supervision management method and system. BACKGROUND

[0002] With the in-depth development of digitalization and intelligentization of industrial manufacturing, the factory production management system gradually realizes real-time monitoring and scheduling of operation process, equipment state and production resources. Based on operation logs, sensor data and task scheduling information, the existing manufacturing execution system, industrial Internet of Things platform and rule-based event processing engine can realize process tracking and abnormal alarm of key processes, providing effective support for ensuring production continuity and efficiency. At the same time, information modeling, graph construction and rule-based monitoring mechanism are also continuously promoting the visualization and standardization of production links.

[0003] However, in actual complex production scenarios, there is a certain rhythm fluctuation and execution path flexibility in operation behavior, and the traditional method still has challenges in capturing nonlinear behavior rhythm pattern, quantifying operation deviation degree and aligning process semantic structure. Especially when considering time dynamics, behavior rhythm and process path semantics and other multi-dimensional factors, how to establish a unified modeling system and realize nested fusion and intelligent analysis of multi-scale data has become one of the important technical problems in the field of production supervision management. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a factory production intelligent supervision management method to solve the problem of how to realize nested fusion and intelligent analysis of multi-scale data.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a factory production intelligent supervision management method, which comprises: constructing a time difference chain based on operation behavior, extracting time interval statistical features and behavior trends within a window through a sliding window, and generating fluctuation feature primitives representing operation rhythm; performing frequency domain transformation on the fluctuation feature primitives, extracting frequency energy distribution features and generating a rhythm fluctuation spectrum;

[0008] Calculating the dynamic time warping distance of the rhythm fluctuation spectrum and the historical rhythm reference spectrum to obtain a spectrum deviation descriptor;

[0009] Assigning a deviation weight value, and nesting and mapping the spectrum deviation descriptor into a standard semantic process flowchart according to the deviation weight value to form a rhythm process nested graph;

[0010] The node-level index semantic vector is extracted from the rhythm process nested graph based on a graph embedding algorithm, and multi-dimensional semantic alignment detection is performed, and an alignment residual metric is output.

[0011] The alignment residual metrics are aggregated to generate a structured misalignment score list, and an abnormal behavior scenario graph is constructed based on a score threshold.

[0012] As a preferred scheme of the factory production intelligent supervision management method, wherein: the spectrum deviation descriptor is used to represent the overall difference degree of the rhythm fluctuation spectrum in the current window compared with the historical rhythm reference spectrum in the main frequency position, the high frequency energy distribution and the spectrum shape characteristics;

[0013] By dynamic time warping algorithm, the frequency components of the rhythm fluctuation spectrum and the historical rhythm reference spectrum are aligned in sections, the cumulative difference value of each corresponding frequency amplitude on the alignment path is calculated, and the deviation feature vector containing time offset compensation is generated as the spectrum deviation descriptor according to the difference distribution.

[0014] As a preferred scheme of the factory production intelligent supervision management method, wherein: the distribution deviation weight value is used to map the spectrum deviation descriptor to the standard semantic process flowchart according to the deviation weight value, forming a rhythm process nested graph, and the specific steps are,

[0015] The node label in the standard semantic process flowchart is parsed, and the process stage identifier associated with the spectrum deviation descriptor is matched;

[0016] The deviation feature vector is distributed to the corresponding node according to the process stage weight, and the deviation value and the timestamp association relationship are recorded in the node attribute, forming a rhythm process nested graph.

[0017] As a preferred scheme of the factory production intelligent supervision management method, wherein: the time difference value chain is composed of the time interval between consecutive operation behaviors, and is marked in combination with the operation behavior category and the station identifier;

[0018] The fluctuation feature base is obtained by setting a fixed time window, sliding extraction of the time difference value chain, and combining the time interval mean, time interval variance, time interval range and difference trend between adjacent windows to generate a structured data unit with rhythm feature expression ability.

[0019] As a preferred scheme of the factory production intelligent supervision management method, wherein: the standard semantic process flowchart is a directed graph structure containing multiple task nodes and connection relationships, and the frequency spectrum vector in the historical rhythm fluctuation spectrum is aggregated by using a clustering method, and then a center spectrum is selected as a representative to obtain.

[0020] As a preferred embodiment of the intelligent supervision and management method for factory production described in this invention, the multidimensional semantic alignment detection compares the semantic vector of the index extracted from each node in the rhythm process nesting diagram with the standard semantic label corresponding to the node to determine the differences between the attributes and output the residual metric value of each node.

[0021] As a preferred embodiment of the intelligent supervision and management method for factory production described in this invention, the structured misalignment scoring list is a distributed structure constructed based on the alignment residual measurement results of each node after aggregation and scoring sorting; the scoring threshold is set according to the historical process execution deviation range, and when the misalignment score exceeds the scoring threshold, the node is identified as an abnormal node and the associated information is used to construct an abnormal behavior scenario diagram.

[0022] Secondly, the present invention provides an intelligent supervision and management system for factory production, including a primitive extraction module, a frequency domain transformation module, a deviation description module, a rhythm embedding module, a semantic alignment module, and a scene construction module;

[0023] The primitive extraction module is used to construct a time difference chain based on the operation behavior, extract the statistical features of the time interval and the behavior trend within the window through a sliding window, and generate a fluctuation feature primitive that represents the operation rhythm; the frequency domain transformation module is used to perform frequency domain transformation on the fluctuation feature primitive, extract the frequency domain energy distribution features and generate a rhythm fluctuation spectrum.

[0024] The deviation description module is used to calculate the dynamic time warping distance between the rhythm fluctuation spectrum and the historical rhythm reference spectrum to obtain the spectrum deviation descriptor.

[0025] The rhythm embedding module is used to assign deviation weight values ​​and to nest and map spectral deviation descriptors to the standard semantic process flow diagram according to the deviation weight values, forming a rhythm process nesting diagram.

[0026] The semantic alignment module is used to extract node-level indicator semantic vectors from the rhythm process nested graph based on the graph embedding algorithm, perform multi-dimensional semantic alignment detection, and output alignment residual metric.

[0027] The scenario construction module is used to aggregate and align residual metrics, generate a structured misalignment score list, and construct an abnormal behavior scenario graph based on the score threshold.

[0028] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the intelligent supervision and management method for factory production as described in the first aspect of the present invention.

[0029] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the intelligent supervision and management method for factory production as described in the first aspect of the present invention.

[0030] The beneficial effects of this invention are as follows: By constructing a time difference value chain and extracting fluctuation feature primitives, combining frequency domain transformation to generate a rhythm fluctuation spectrum, and calculating its dynamic time warping distance with the historical rhythm reference spectrum, it is possible to model and quantify the deviation of operational behavior rhythm, identify rhythm abrupt changes and frequency anomalies, and improve the sensitivity and recognition accuracy of potential operational anomalies. By nesting and mapping spectral deviation descriptors to a standard semantic process flow diagram, and extracting node-level indicator semantic vectors based on graph embedding algorithms, multi-dimensional semantic alignment detection is performed to achieve structured comparison and deviation localization between operational behavior and process semantics, solving the problem of difficulty in semantically interpreting and tracing the source of operational behavior anomalies in the process path. This enhances the ability to identify anomalies and semantic misalignments during production supervision, and improves the depth of perception of dynamic behavior and the structured expression ability of anomaly diagnosis. Attached Figure Description

[0031] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Fig. 1 Overall flowchart of intelligent supervision and management methods for factory production.

[0033] Fig. 2 A schematic diagram of rhythm wave spectrum generation.

[0034] Fig. 3 This is a schematic diagram of the nested process of rhythm technology.

[0035] Fig. 4 Build a flowchart for the abnormal behavior scenario diagram. Detailed Implementation

[0036] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0037] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0038] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0039] Reference Figs. 1-4 This is one embodiment of the present invention, which provides a method for intelligent supervision and management of factory production, including the following steps:

[0040] S1: Construct a time difference chain based on operational behavior, extract the statistical features and behavioral trends of time intervals within the window through a sliding window, and generate fluctuation feature primitives that represent the rhythm of operations.

[0041] Specifically, it includes the following steps:

[0042] S1.1: Record the operational behaviors during the production process to form a sequence of operational behaviors.

[0043] Specifically, the operation behavior record includes the operation occurrence timestamp, operation behavior type identifier, and workstation number. The operation behavior type identifier is used to distinguish specific operation actions, such as tightening, handling, inspection, and confirmation; the workstation number is used to specify the physical location where the operation occurred. All records are arranged in ascending order of timestamps, forming an operation behavior sequence.

[0044] S1.2: Calculate the time difference between adjacent operations based on the sequence of operations, and combine the operation types to form a time difference value chain.

[0045] Specifically, based on the timestamps of adjacent operations in the sequence of operations, the time difference between adjacent operations is calculated in seconds, forming the original time interval sequence. For each pair of operations, the operation type identifier (e.g., handling → tightening) and the corresponding workstation number are appended to the time difference, forming a time difference data unit. All time difference data units are arranged in chronological order to form a time difference chain.

[0046] S1.3: Perform sliding window extraction on the time difference chain to extract continuous segments from the time difference chain.

[0047] Specifically, a fixed time window length and sliding step size are set to segment the time difference chain. The sliding window slides along the time difference chain in chronological order, extracting a fixed-length continuous time difference sequence as a window segment each time. For example, if the time window length is set to 5 and the sliding step size is set to 1, then each window contains 5 consecutive time difference data units. The sliding windows can overlap to cover the entire time difference chain, generating a set of window segments.

[0048] S1.4: Perform statistical calculations on continuous segments in the time difference chain to obtain statistical indicators used to express the characteristics of the operation rhythm.

[0049] Specifically, statistical analysis is performed on the time difference data units in each sliding window to extract statistical indicators. These indicators include the mean of the time interval, the variance of the time interval, the range of the time interval, the diversity of operation behavior types within the window, and the trend of the mean change before and after the window.

[0050] It should be noted that the time interval range refers to the difference between the maximum and minimum time intervals. The diversity of operation types within the window refers to the number of different combinations of operation behaviors. The trend of the mean change before and after the window refers to the difference between the mean of the previous window and the mean of the current window.

[0051] Statistical indicators are used to reflect the stability, volatility, and trend changes of operational behavior.

[0052] S1.5: Combine statistical indicators into structured data units and combine them with operational behavior types to form fluctuation characteristic primitives.

[0053] Specifically, statistical indicators are combined sequentially into structured feature vectors. These structured feature vectors are then combined with the most frequently occurring operation behavior type label in the current sliding window to generate fluctuation feature primitives. Each fluctuation feature primitive corresponds to a sliding window, possessing a fixed-length structure, behavior type identifier, and rhythm fluctuation characteristics, which are used in subsequent frequency domain analysis steps to generate rhythm fluctuation spectra.

[0054] Ideally, by recording and labeling operational behaviors during the production process with timestamps and behavior types, and constructing a time difference chain, the temporal rhythm relationship between consecutive operations can be accurately captured. A sliding window mechanism is used to extract time difference sequences in segments, and statistical indicators such as the mean, variance, range, operational behavior diversity, and mean change trend of time intervals are calculated. This comprehensively characterizes the temporal features and rhythmic fluctuations of operational behaviors. Encoding these statistical indicators into structured feature vectors and combining them with major operational behavior types to generate fluctuation feature primitives allows the rhythmic characteristics of operational behaviors to be input into subsequent frequency domain processing in a unified form. This enables the modeling and quantification of irregular operational rhythms, laying the foundation for identifying potential rhythmic anomalies in the production process and enhancing sensitivity and expressive power in complex behavioral scenarios.

[0055] S2: Perform frequency domain transformation on the wave feature primitives, extract the frequency domain energy distribution features, and generate the rhythm wave spectrum.

[0056] Specifically, it includes the following steps:

[0057] S2.1: Arrange the wave characteristic primitives in chronological order to form a wave characteristic primitive sequence indexed by equal time intervals.

[0058] Specifically, the fluctuation feature primitives are arranged in ascending order of timestamps. Each fluctuation feature primitive contains a structured feature vector, which is derived from the rhythm statistics extracted by the sliding window. The time index is determined according to the start time of the sliding window, ensuring that the fluctuation feature primitives have continuity and equal intervals on the time axis.

[0059] S2.2: Normalize the statistical indicators in the fluctuation characteristic primitive sequence.

[0060] Specifically, for each statistical indicator, the maximum and minimum values ​​within the entire sequence are extracted, and the original indicator values ​​are linearly mapped to the interval [0,1] using the maximum-minimum normalization method. Normalization improves the scale consistency of statistical indicators across dimensions and avoids interference caused by differences in dimensions during subsequent frequency domain transformation.

[0061] S2.3: The frequency domain transformation of the wave characteristic primitive sequence is performed by the Fast Fourier Transform to obtain the spectral representation.

[0062] Specifically, the frequency domain transformation employs the Fast Fourier Transform (FFT) method to map the statistical index sequence in the time domain to a complex spectrum in the frequency domain. The sampling frequency is set based on the time index of the wave characteristic primitive sequence, and the sampling rate is calculated based on the sampling frequency. A Hamming window function is used to reduce spectral leakage.

[0063] The normalized wave characteristic primitive sequence is multiplied by a Hamming window function for windowing, and zero-padding and FFT calculations are performed to obtain the amplitude spectrum, power spectrum and frequency axis.

[0064] S2.4: Extract frequency domain features based on spectral representation to construct rhythmic wave spectrum.

[0065] Specifically, frequency domain characteristics include dominant frequency, high-frequency energy proportion, spectral kurtosis, spectral variance, and frequency band energy distribution.

[0066] The dominant frequency is obtained by selecting the frequency corresponding to the largest amplitude value in the amplitude spectrum.

[0067] The high-frequency energy proportion was obtained by setting a high-frequency threshold of 0.3 times the sampling rate and calculating the proportion of the total power spectrum exceeding the high-frequency threshold to the total power. The high-frequency threshold was determined empirically and adjusted experimentally.

[0068] Spectral kurtosis measures the sharpness of spectral peaks, while spectral variance reflects the dispersion of energy distribution.

[0069] Frequency band energy distribution refers to dividing the entire spectrum into low-frequency, mid-frequency, and high-frequency bands, and calculating the energy proportion within each band. The spectrum division is based on the physical periodic characteristics of operational behavior and the spectral analysis requirements of signal processing. The low-frequency band is 0-0.01Hz, the mid-frequency band is 0.01-0.1Hz, and the high-frequency band is 0.1-0.2Hz. The energy proportion of each band is obtained by summing the power spectra at all frequency points within the band and then dividing by the total power of the entire spectrum.

[0070] All frequency domain features are structured into a JSON format file and saved as a rhythmic wave spectrum.

[0071] Preferably, by constructing the fluctuation feature primitives into a sequence of fluctuation feature primitives with equally spaced time indices and normalizing the fluctuation feature primitive sequence, the consistency and effectiveness of rhythm statistical indicators in various dimensions in frequency domain analysis are ensured. By using Fast Fourier Transform combined with Hamming window function and zero-padding operation, the rhythm signal of the operation behavior can be accurately converted from the time domain to the frequency domain, effectively suppressing spectral leakage and improving spectral clarity. In the frequency domain, multi-dimensional features such as the main frequency, high-frequency energy ratio, spectral kurtosis, spectral variance, and frequency band energy distribution are extracted to construct a rhythm fluctuation spectrum. This not only reveals the structural features of the rhythm of the operation behavior in fine granularity, but also sensitively captures hidden abnormal fluctuation patterns, thus providing a highly expressive input basis for subsequent spectral deviation identification and rhythm anomaly analysis.

[0072] S3: Calculate the dynamic time warping distance between the rhythm fluctuation spectrum and the historical rhythm reference spectrum to obtain the spectrum deviation descriptor.

[0073] Specifically, it includes the following steps:

[0074] S3.1: Extract frequency component sequences from the rhythm fluctuation spectrum and the historical rhythm reference spectrum, respectively.

[0075] It should be noted that the historical rhythm reference spectrum is calculated from the database based on historical operational behavior.

[0076] Structured spectral data is extracted from the rhythm fluctuation spectrum and historical rhythm reference spectrum. Each structured spectral data point contains a frequency axis and corresponding power values, amplitudes, and other frequency domain features. Frequency components on the frequency axis are extracted as the main analytical line, forming a frequency component sequence. The frequency component sequence includes information such as the dominant frequency, high-frequency energy distribution points, and characteristic frequency points, which are used for subsequent alignment processing.

[0077] S3.2: The frequency component sequence of the rhythm fluctuation spectrum and the frequency component sequence of the historical rhythm reference spectrum are used as inputs, and the dynamic time warping algorithm is used to perform segment-by-segment alignment processing.

[0078] Specifically, the Dynamic Time Warping (DTW) algorithm is selected as the core alignment method. The frequency component sequence of the rhythm fluctuation spectrum is taken as one-dimensional time series input, and the frequency component sequence of the historical rhythm reference spectrum is taken as another sequence input, constructing a dynamic time warping distance matrix. The DTW algorithm finds the shortest path between the two frequency component sequences to minimize the overall matching distance. The position index and corresponding amplitude information of each matched frequency point pair are retained in the original sequence for subsequent error calculation.

[0079] S3.3: After the alignment is completed by the dynamic time warping algorithm, obtain the amplitude difference between each corresponding frequency component pair on the alignment path, and perform cumulative calculation to form a total distance value representing the overall degree of spectral difference.

[0080] Specifically, for each pair of matched frequency points in the DTW alignment path, the absolute difference in amplitude between the pair of frequency points is calculated. All differences are accumulated and summed in alignment order to obtain a scalar total distance value. For example, if the alignment path contains 80 pairs of frequency points, and the amplitude difference between each pair is 0.01, 0.03, 0.04, ..., then these 80 values ​​are accumulated to obtain the offset measure of the rhythm fluctuation spectrum relative to the historical rhythm reference spectrum in the overall frequency domain structure as the total distance value.

[0081] S3.4: Based on the total distance value, a combined feature vector is constructed by combining the dynamic time warping path, and the output is a spectral deviation descriptor.

[0082] Specifically, while obtaining the total distance value, information from the alignment path is extracted as supplementary features to enrich the spectral deviation descriptor structure. Supplementary features include, but are not limited to: alignment path length (i.e., the number of matched frequency point pairs), maximum frequency axis offset (i.e., the maximum jump between frequency indices), standard deviation of amplitude difference (measuring the dispersion of the difference distribution), and the number of high amplitude offset segments (e.g., the number of frequency difference segments that continuously exceed a certain threshold).

[0083] The total distance value and supplementary features are concatenated into a one-dimensional structured vector to form the spectral deviation descriptor. This spectral deviation descriptor, as a representation of the difference between the rhythmic fluctuation spectrum and the historical rhythmic reference spectrum, is used in subsequent nested mapping steps to complete the alignment analysis of the semantic process.

[0084] Ideally, by performing fine-grained alignment and quantitative analysis of the frequency domain differences between the rhythm fluctuation spectrum and the historical rhythm reference spectrum, highly sensitive identification of rhythm changes in production operation behavior is achieved. Processing the frequency component sequence using a dynamic time warping algorithm not only effectively eliminates the effects of nonlinear mismatch in frequency domain changes but also preserves the positional information and amplitude characteristics of key frequency points. Furthermore, by accumulating amplitude differences and statistically analyzing multidimensional alignment paths, a spectral deviation descriptor is constructed. This not only provides a quantitative indicator of the overall rhythm shift but also captures the local structural features of rhythm disturbances, helping to improve the accuracy and robustness of the method in abnormal rhythm detection, semantic behavior mapping, and process monitoring.

[0085] S4: Assign deviation weight values ​​and map the spectral deviation descriptors to the standard semantic process flow diagram according to the deviation weight values ​​to form a rhythm process nesting diagram.

[0086] Specifically, it includes the following steps:

[0087] S4.1: Parse the node labels in the standard semantic process flow diagram and extract the process stage identifiers and node attributes.

[0088] It should be noted that the standard semantic process flow diagram uses nodes to represent process stages and directed edges to represent the dependency order between process flows.

[0089] All nodes in the standard semantic process flow diagram are parsed sequentially to extract the process stage identifier and basic attribute information of each node, including the node start timestamp, end timestamp, workstation number, and behavior label. The node time information is used to determine the correspondence of the spectral deviation descriptor on the time axis, and the node behavior label is used to assist in matching the operation type.

[0090] S4.2: Perform time-process mapping on the spectral deviation descriptor.

[0091] Specifically, the spectral deviation descriptor originates from the alignment process between the rhythmic fluctuation spectrum and the historical rhythmic reference spectrum, retaining a time index. The time index in the spectral deviation descriptor is compared with the node time range in the standard semantic process flow diagram to identify the covered node intervals.

[0092] The mapping method is as follows: if the time index of the spectral deviation descriptor falls within the start and end time range of a certain node, then the spectral deviation descriptor is mapped to the process stage represented by that node. If it spans multiple nodes, the overlap ratio of the spectral deviation descriptor in the time ranges of different nodes is calculated to form a deviation weight value.

[0093] S4.3: Assign deviation weight values ​​according to the duration of the process stage.

[0094] Specifically, for a spectral deviation descriptor spanning multiple process stages, a linear allocation is performed based on the time proportion in each process stage. For example, if a spectral deviation descriptor has a total duration of 60 seconds, spanning two process stages A (30 seconds) and B (30 seconds), then 50% deviation weight is allocated to stage A and stage B respectively. If a process stage only contains a portion of the spectral deviation descriptor, the proportion is calculated based on the actual coverage time and weighted accordingly. The allocation result is represented as a set of key-value pairs <process stage identifier, deviation weight value>.

[0095] S4.4: Write the spectral deviation descriptor into the corresponding node according to the deviation weight value, and obtain the node set.

[0096] Specifically, the total distance value and supplementary features of each spectral deviation descriptor are associated with the corresponding process stage identifier and scaled according to the deviation weight value. For example, if the total distance value of the spectral deviation descriptor is 0.48 and the deviation weight of a certain node is 0.3, then the deviation value written to that node is 0.48 × 0.3 = 0.144. All spectral deviation descriptors are written to the nodes in the flowchart one by one in the above manner. Finally, each node contains a subset of spectral deviation features, forming a node set.

[0097] S4.5: Calculate the weights of directed edges based on process dependencies and obtain the edge set.

[0098] Specifically, directed edges in a standard semantic process flow diagram reflect the logical or physical sequence between operational process stages. For any pair of adjacent nodes with a dependency relationship, a subset of spectral deviation descriptions is extracted from the node set. The weight of the directed edge is obtained by calculating the absolute value of the difference between the deviation values ​​of the pair of nodes. For example, if the deviation value of the preceding node is 0.144 and the deviation value of the following node is 0.217, then the edge weight is |0.217-0.144|=0.073. This edge weight represents the degree of shift in rhythm anomaly from the previous stage to the next stage, and is interpretable.

[0099] S4.6: Construct a nested graph of rhythmic processes based on the set of nodes and the set of edges.

[0100] Preferably, the set of all nodes containing spectral deviation features and the set of edges calculated based on process flow dependencies are jointly represented as a graph structure. Each node in the graph structure has a spectral deviation feature vector as an attribute, and each edge has a weight value. The graph structure is a rhythmic process nested graph, reflecting the distribution and propagation trend of anomalies at the rhythm fluctuation level in different process stages, and can be used for subsequent tasks such as anomaly identification, rhythm optimization, or process reconstruction.

[0101] Preferably, by nesting and mapping spectral deviation descriptors with standard semantic process flow diagrams, a rhythmic process nesting diagram is formed, achieving precise alignment between rhythmic fluctuation information and the process flow. This allows rhythmic anomalies to be located at the process node level with fine granularity. The structured expression based on node deviation values ​​and edge weights not only clarifies the spatial propagation path of anomalies in the process flow but also enhances the interpretability and traceability of the rhythmic anomaly formation mechanism. The constructed rhythmic process nesting diagram provides a data foundation and structural support for subsequent anomaly identification, process optimization, and decision support, improving the intelligence and refinement of factory production management.

[0102] S5: Based on the graph embedding algorithm, extract node-level indicator semantic vectors from the rhythm process nested graph, perform multi-dimensional semantic alignment detection, and output alignment residual metric.

[0103] Specifically, it includes the following steps:

[0104] S5.1: Extract node attribute data from the rhythm process nested graph and construct a node feature matrix.

[0105] Specifically, each node in the rhythmic process nested graph contains a subset of spectral deviation features, including multiple dimensions such as total distance value, alignment path length, maximum frequency axis offset, amplitude difference standard deviation, and number of high amplitude offset segments. Each of these spectral deviation feature dimensions is extracted and arranged in a uniform order to form a node feature vector. Using all nodes in the graph as samples, and each node's feature vector as a row, a two-dimensional node feature matrix is ​​constructed. For example, if the graph contains 30 nodes, and each node's spectral deviation feature is 5-dimensional, then the node feature matrix size is 30×5.

[0106] S5.2: A two-layer graph convolutional network is used for feature propagation to generate node-level indicator semantic vectors.

[0107] Specifically, the graph structure information based on the rhythm process nested graph is combined with the node feature matrix and input into the graph convolutional network for feature propagation.

[0108] The graph convolutional network adopts a two-layer stacked structure: the first layer of graph convolution performs linear transformation and weighted aggregation on the initial node features and first-order neighbor features to generate intermediate semantic representations;

[0109] The second-layer graph convolution continues to propagate to a wider neighborhood based on the intermediate semantic representation, outputting the final node-level index semantic vector.

[0110] Each graph convolutional layer performs ReLU activation and uses a normalized adjacency matrix for structural weighting. For example, if the input node feature dimension is 5, the first layer output dimension is 16, and the second layer output dimension is 8, then each node will ultimately generate an 8-dimensional node-level index semantic vector, representing the embedded semantic expression of the node under rhythmic fluctuations deviating from the structure.

[0111] S5.3: Obtain the standard semantic label vector for each node from the standard semantic process flow diagram.

[0112] Specifically, each node in the standard semantic process flow diagram has a process stage identifier and node attribute labels, including behavior type labels, workstation number, operation duration, and historical expected rhythm intervals. The aforementioned label data for each node is vectorized to form a structured standard semantic label vector.

[0113] The encoding method includes: one-hot encoding for behavior types, ordinal encoding for workstation numbers, and normalized numerical representation for operation duration and rhythm intervals. The encoding results of each node are concatenated in a fixed order to form a standard semantic tag vector.

[0114] S5.4: Perform multi-dimensional semantic alignment detection on node-level indicator semantic vectors and standard semantic label vectors, perform cosine similarity comparison and Euclidean distance residual calculation, and generate an alignment residual metric set.

[0115] Specifically, for each node in the rhythm process nesting diagram, the corresponding node-level indicator semantic vector is compared with the standard semantic label vector of the node with the same identifier in the standard semantic process flow diagram. The comparison methods include two indicators: cosine similarity comparison and Euclidean distance residual calculation.

[0116] Cosine similarity comparison: Calculate the cosine of the angle between two vectors, with a value range of [-1, 1], and use it to measure semantic direction consistency.

[0117] Euclidean distance residual calculation: Calculates the straight-line distance between two vectors in Euclidean space, representing the overall residual degree of semantic expression.

[0118] The cosine similarity comparison is used as the output of the alignment metric, and the Euclidean distance residual calculation is used as the output of the residual metric. Both are saved as structured index records. The cosine similarity and Euclidean residual results of all nodes form the alignment residual metric set.

[0119] Preferably, by performing a graph-structured nested comparison between spectral deviation features and standard semantic labels, and leveraging a graph convolutional network, a multi-dimensional semantic alignment mechanism is constructed between the rhythmic process nested graph and the standard semantic process flow diagram. Through a dual comparison method using cosine similarity and Euclidean distance, both the consistency of node semantics can be quantified, and the residual shift of node semantics can be accurately assessed. This effectively enables semantic diagnosis and quantitative expression of production rhythm anomalies, improving the intelligence, refinement, and interpretability of production process supervision and management.

[0120] S6: Aggregate and align residual metrics to generate a structured misalignment score list, and construct an abnormal behavior scenario graph based on the score threshold.

[0121] Specifically, it includes the following steps:

[0122] S6.1: Calculate the misalignment score for each node based on the alignment residual metric set.

[0123] Specifically, the cosine similarity in the alignment residual metric set is transformed by cosine difference, and the Euclidean distance residual is normalized.

[0124] Using the Euclidean distance residual as the basic linear term, and adding cosine similarity as a weighting factor, the misalignment score calculation expression is as follows:

[0125] S=α×D cos +β×R norm ;

[0126] Where S represents misplaced rating, and D... cos R represents the cosine similarity after cosine difference transformation. norm The normalized Euclidean distance residual is denoted by α, which is the cosine similarity ratio coefficient, and β is the Euclidean distance residual ratio coefficient. α and β are initially set to 0.5 and are adjusted according to the proportion of directional deviation and magnitude shift, satisfying α + β = 1.

[0127] S6.2: Generate a structured list of misaligned scores based on misaligned scores.

[0128] It should be noted that the structured misaligned scoring list is arranged in descending order of misaligned scores, while nodes with the same score are arranged in ascending order of timestamps.

[0129] S6.3: Set the scoring threshold based on the distribution range of node misalignment scores in the historical process flow.

[0130] Specifically, multiple sets of misaligned score lists are extracted from the historical semantic process flow, the historical score mean and historical score standard deviation are calculated, and the Z-Score method is used to set the score threshold as the historical score mean + 1.5 × historical score standard deviation.

[0131] S6.4: Construct an abnormal behavior scenario graph based on the scoring threshold.

[0132] Specifically, iterate through the list of misaligned scores and mark nodes whose misaligned scores exceed the scoring threshold as anomalous nodes. Extract all anomalous nodes and integrate them into a set of anomalous nodes.

[0133] Based on the dependency edges in the standard semantic flowchart, only the edges connecting two abnormal nodes are retained to obtain the set of abnormal edges.

[0134] An abnormal behavior scenario graph is constructed based on the abnormal node set and abnormal edge set, and the graph is drawn using the NetworkX library in Python to achieve visualization support.

[0135] Ideally, by constructing a structured misalignment scoring list and forming an abnormal behavior scenario diagram, quantitative identification and centralized presentation of operational rhythm deviations and semantic misalignments are achieved. Combining cosine similarity and Euclidean distance residuals to comprehensively score node alignment errors improves the robustness and multi-dimensional expressive power of misalignment assessment. Through statistical scoring threshold settings, the system dynamically adapts to the anomaly detection needs under different process backgrounds. The abnormal behavior scenario diagram constructed based on highly misaligned nodes not only clarifies the structural distribution of abnormal behaviors but also supports intuitive localization and correlation analysis of out-of-control links, thereby improving the accuracy and interpretability of anomaly identification in complex process scenarios.

[0136] This embodiment also provides a factory production intelligent supervision and management system, including: a primitive extraction module, a frequency domain transformation module, a deviation description module, a rhythm embedding module, a semantic alignment module, and a scene construction module;

[0137] The primitive extraction module is used to construct a time difference chain based on operational behavior. It extracts the statistical features and behavioral trends of the time interval within a sliding window to generate fluctuation feature primitives that characterize the rhythm of the operation.

[0138] The frequency domain transformation module is used to perform frequency domain transformation on the wave feature primitives, extract the frequency domain energy distribution features, and generate the rhythm wave spectrum.

[0139] The deviation description module is used to calculate the dynamic time warp distance between the rhythm fluctuation spectrum and the historical rhythm reference spectrum to obtain the spectrum deviation descriptor;

[0140] The rhythm embedding module is used to assign deviation weight values ​​and to map the spectral deviation descriptors to the standard semantic process flow diagram according to the deviation weight values, forming a rhythm process nesting diagram.

[0141] The semantic alignment module is used to extract node-level indicator semantic vectors from the rhythm process nested graph based on the graph embedding algorithm, perform multi-dimensional semantic alignment detection, and output alignment residual metric.

[0142] The scenario building module is used to aggregate and align residual metrics, generate a structured misalignment score list, and construct an abnormal behavior scenario graph based on the score threshold.

[0143] This embodiment also provides a computer device applicable to the intelligent supervision and management method for factory production, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the intelligent supervision and management method for factory production as proposed in the above embodiment.

[0144] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0145] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the intelligent supervision and management method for factory production as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0146] In summary, this invention achieves modeling and deviation quantification of operational rhythm by constructing a time difference value chain and extracting fluctuation feature primitives, generating a rhythm fluctuation spectrum through frequency domain transformation, and calculating the dynamic time warping distance between the rhythm fluctuation spectrum and the historical rhythm reference spectrum. This enables the identification of rhythmic abrupt changes and frequency anomalies, improving the sensitivity and accuracy of potential operational anomalies. By nesting and mapping spectral deviation descriptors to a standard semantic process flow diagram, and extracting node-level indicator semantic vectors based on graph embedding algorithms, multi-dimensional semantic alignment detection is performed to achieve structured comparison and deviation localization between operational behavior and process semantics. This solves the problem of difficulty in semantically interpreting and tracing the source of operational behavior anomalies in the process path. It enhances the ability to identify anomalies and semantic misalignments during production supervision, improves the depth of perception of dynamic behavior, and enhances the structured expression ability for anomaly diagnosis.

[0147] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent supervision and management of factory production, characterized in that: include, A time difference chain is constructed based on operational behavior. The statistical features and behavioral trends of the time interval within the window are extracted through a sliding window to generate fluctuation feature primitives that characterize the rhythm of the operation. Frequency domain transformation is performed on the wave feature primitives to extract frequency domain energy distribution features and generate rhythm wave spectrum; The dynamic time warping distance between the rhythm fluctuation spectrum and the historical rhythm reference spectrum is calculated to obtain the spectrum deviation descriptor; Assign deviation weight values ​​and map the spectral deviation descriptors to the standard semantic process flow diagram according to the nested deviation weight values ​​to form a rhythmic process nesting diagram. Based on the graph embedding algorithm, node-level indicator semantic vectors are extracted from the rhythm process nested graph, and multi-dimensional semantic alignment detection is performed to output alignment residual metric. Aggregate and align residual metrics to generate a structured misalignment score list, and construct an abnormal behavior scenario graph based on the score threshold; The spectral deviation descriptor is used to characterize the overall difference between the rhythm fluctuation spectrum within the current window and the historical rhythm reference spectrum in terms of dominant frequency position, high-frequency energy distribution, and spectral morphology characteristics. The frequency components of the rhythm fluctuation spectrum and the historical rhythm reference spectrum are aligned segment by segment using a dynamic time warping algorithm. The cumulative difference value of the corresponding frequency amplitudes on the alignment path is calculated, and a deviation feature vector containing time offset compensation is generated as a spectrum deviation descriptor based on the difference distribution.

2. The intelligent supervision and management method for factory production as described in claim 1, characterized in that: The allocation of deviation weight values ​​involves nesting and mapping the spectral deviation descriptors to the standard semantic process flow diagram according to the deviation weight values, forming a rhythmic process nesting diagram. The specific steps are as follows: Parse the node labels in the standard semantic process flow diagram and match the process stage identifiers associated with the spectral deviation descriptor; The deviation feature vectors are assigned to the corresponding nodes according to the weight of the process stage, and the deviation value and timestamp relationship are recorded in the node attributes to form a rhythm process nesting diagram.

3. The intelligent supervision and management method for factory production as described in claim 1, characterized in that: The time difference value chain consists of the time interval between consecutive operation behaviors, and is marked in combination with the operation behavior category and workstation identifier; The fluctuation feature primitives extract the time difference chain by setting a fixed time window, and generate structured data units by combining the mean of the time interval, the variance of the time interval, the range of the time interval, and the difference trend between adjacent windows within the window.

4. The intelligent supervision and management method for factory production as described in claim 1, characterized in that: The standard semantic process flow diagram is a directed graph structure containing multiple task nodes and connection relationships. It is obtained by aggregating the spectral vectors in the historical rhythm fluctuation spectrum using a clustering method, and then selecting the central spectrum as a representative.

5. The intelligent supervision and management method for factory production as described in claim 1, characterized in that: The multidimensional semantic alignment detection compares the semantic vector of the index extracted from each node in the rhythm process nested graph with the standard semantic label corresponding to the node to determine the differences between the attributes and output the residual metric value of each node.

6. The intelligent supervision and management method for factory production as described in claim 1, characterized in that: The structured misalignment scoring list is a distributed structure built based on the alignment residual measurement results of each node, after aggregation and scoring sorting; the scoring threshold is set according to the historical process execution deviation range. When the misalignment score exceeds the scoring threshold, the node is marked as an abnormal node and the associated information is used to construct an abnormal behavior scenario graph.

7. A factory production intelligent supervision and management system, characterized in that: It includes a primitive extraction module, a frequency domain transformation module, an offset description module, a rhythm embedding module, a semantic alignment module, and a scene construction module; The primitive extraction module is used to construct a time difference chain based on the operation behavior, extract the statistical features of the time interval and the behavioral trend within the window through a sliding window, and generate a fluctuation feature primitive that represents the rhythm of the operation. The frequency domain transformation module is used to perform frequency domain transformation on the wave feature primitives, extract frequency domain energy distribution features, and generate rhythm wave spectrum; The deviation description module is used to calculate the dynamic time warping distance between the rhythm fluctuation spectrum and the historical rhythm reference spectrum to obtain the spectrum deviation descriptor. The rhythm embedding module is used to assign deviation weight values ​​and to nest and map the spectral deviation descriptors to the standard semantic process flow diagram according to the deviation weight values, thereby forming a rhythm process nesting diagram. The semantic alignment module is used to extract node-level indicator semantic vectors from the rhythm process nested graph based on the graph embedding algorithm, perform multi-dimensional semantic alignment detection, and output alignment residual metric. The scenario construction module is used to aggregate and align residual metrics, generate a structured misalignment score list, and construct an abnormal behavior scenario graph based on the score threshold. The spectral deviation descriptor is used to characterize the overall difference between the rhythm fluctuation spectrum within the current window and the historical rhythm reference spectrum in terms of dominant frequency position, high-frequency energy distribution, and spectral morphology characteristics. The frequency components of the rhythm fluctuation spectrum and the historical rhythm reference spectrum are aligned segment by segment using a dynamic time warping algorithm. The cumulative difference value of the corresponding frequency amplitudes on the alignment path is calculated, and a deviation feature vector containing time offset compensation is generated as a spectrum deviation descriptor based on the difference distribution.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent supervision and management method for factory production as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent supervision and management method for factory production as described in any one of claims 1 to 6.

Images