Fried instant noodle production line fault monitoring and remote early warning method and system

Abstract

The invention provides a fault monitoring and remote early warning method and system for a fried instant noodle production line, and relates to the technical field of fault monitoring. Synchronously activating a sensing layer to obtain a process node data set after the production line runs; matching features are extracted through a process chain, an anomaly recognition library is called, an anomaly sensor mapped with process nodes is established, the anomaly sensor is used for carrying out process recognition on a corresponding process node data set, and an independent anomaly recognition result is established; and performing cross-process influence analysis by using a state propagation function, establishing a collaborative anomaly recognition result to compensate an independent anomaly recognition result, and reporting an anomaly early warning. According to the invention, the technical problems of low fault identification accuracy and delayed early warning response caused by isolated monitoring of each process and lack of linkage analysis in the prior art are solved, and the technical effects of improving the fault identification accuracy and early warning real-time performance and realizing cooperative fault monitoring among the processes are achieved.

Description

Technical Field

[0001] This invention relates to the field of fault monitoring technology, specifically to a method and system for fault monitoring and remote early warning of instant noodle production lines. Background Technology

[0002] As a popular ready-to-eat food, fried instant noodles rely heavily on automated assembly line operations for production. To ensure product quality and production efficiency, monitoring the operational status and providing early warnings for faults in fried instant noodle production lines are crucial aspects of production management. Existing fault monitoring methods for fried instant noodle production lines often involve installing sensors on key equipment to collect parameters such as temperature, vibration, and current to assess the equipment's operating status and trigger alarms when thresholds are exceeded. These methods mostly focus on monitoring single equipment or specific sections, lacking unified modeling and collaborative analysis of the entire production line's process chain. This results in fragmented fault identification, an inability to accurately reflect the transmission relationship between processes, difficulty in precisely locating the root cause of faults, and delayed early warning responses, failing to provide timely and effective fault handling information to production personnel. Consequently, production downtime is prolonged, and product quality suffers. Summary of the Invention

[0003] This invention provides a method and system for fault monitoring and remote early warning of a fried instant noodle production line. It solves the technical problems of low fault identification accuracy and delayed early warning response caused by isolated monitoring of each process and lack of inter-process linkage analysis in the existing technology. It achieves the technical effect of improving fault identification accuracy and real-time early warning, and realizing collaborative fault monitoring between processes.

[0004] In view of the above problems, on the one hand, the present invention provides a method for fault monitoring and remote early warning of a fried instant noodle production line. The method includes: after reading the process chain of the fried instant noodle production line, configuring a perception layer at each process node; when the fried instant noodle production line is running, synchronously activating the perception layer to perform production data perception of the process nodes and establishing a process node dataset; extracting matching features from the process chain, calling an anomaly identification library according to the matching features, and establishing an anomaly sensor mapped to the process nodes using the anomaly identification library; using the anomaly sensor to perform process identification of the corresponding process node dataset and establishing an independent anomaly identification result; after extracting the process node dataset into a process state vector, using a state propagation function to perform cross-process influence analysis of the process state vector and establishing a collaborative anomaly identification result; after compensating the independent anomaly identification result with the collaborative anomaly identification result, reporting an anomaly warning.

[0005] Preferably, the step of extracting the process node dataset into process state vectors and then using a state propagation function to perform cross-process influence analysis of the process state vectors and establish collaborative anomaly identification results includes: treating each process node of the instant noodle production line as a local state field source point, performing time evolution based on the process state vectors to establish a dynamic state field; constructing a state coupling path based on the dynamic state field, then using the state propagation function to perform inter-process linkage strength and state influence analysis to construct a third-order collaborative tensor; and using the third-order collaborative tensor to perform collaborative anomaly tracing and establish collaborative anomaly identification results.

[0006] Preferably, after constructing the state coupling path based on the dynamic state field, the step of using the state propagation function to analyze the linkage strength and state influence between processes and constructing a third-order collaborative tensor includes: configuring the propagation focus point using the independent anomaly identification result and the state coupling path; using the propagation focus point as the search starting point, performing a backtracking fluctuation search of the state coupling path and establishing a backtracking fluctuation search result; after reconstructing the search starting point based on the backtracking fluctuation search result, using the state propagation function to analyze the linkage strength and state influence between processes and construct a third-order collaborative tensor.

[0007] Preferably, the step of extracting matching features from the process chain, calling an anomaly recognition library based on the matching features, and establishing an anomaly sensor mapped to the process nodes using the anomaly recognition library includes: establishing a process chain sequence based on the process chain, wherein the processes in the process chain sequence include cutting, steaming, shaping, frying, cooling, and packaging; extracting feature tuples from each process in the process chain sequence, and then extracting global features of the process chain sequence; and using the feature tuples and the global features as matching features to call the anomaly recognition library to establish the anomaly sensor.

[0008] Preferably, the sensing layer includes a basic sensing unit, a noise filtering unit, and a multimodal data fusion interface. The sensing parameters of the basic sensing unit include temperature, humidity, pressure, vibration, sound, visual image, flow rate, and current. The noise filtering unit is used for data preprocessing of the basic sensing unit, and the multimodal data fusion interface is used for aligning and fusing multiple types of sensing data.

[0009] Preferably, the step of reporting an anomaly warning after compensating the independent anomaly identification result with the collaborative anomaly identification result includes: generating an anomaly status identifier for each process node based on the independent anomaly identification result; performing source tracing compensation of the anomaly status identifier using the collaborative anomaly identification result; and reporting an anomaly warning based on the source tracing compensation of the anomaly status identifier.

[0010] Preferably, after reporting the abnormal warning, the process includes: obtaining the warning response of the fried instant noodle production line, packaging and recording the warning response and the abnormal warning as abnormal historical data; calling the abnormal historical data to identify the sensitive features of the fried instant noodle production line, and using the sensitive features to monitor and manage the fried instant noodles.

[0011] Preferably, the establishment of collaborative anomaly identification results further includes: introducing an oil product state sensor into the perception layer to establish an oil product monitoring time series; using the oil product monitoring time series to fit oil product deterioration and establish a time-series evolution anomaly; constructing a time-series risk fluctuation path through the time-series evolution anomaly and the state coupling path; and using the time-series risk fluctuation path to compensate for the collaborative anomaly identification results.

[0012] Preferably, after reporting the abnormal warning, the process includes: creating a warning verification window based on the abnormal warning; determining whether there is a production line maintenance response in the warning verification window; if there is no production line maintenance response, then performing a shutdown process on the instant noodle frying production line.

[0013] On the other hand, the present invention also provides a fault monitoring and remote early warning system for a fried instant noodle production line. The system includes: a perception configuration module, used to configure a perception layer at each process node after reading the process chain of the fried instant noodle production line; a production data perception module, used to synchronously activate the perception layer to perform production data perception of the process nodes after the fried instant noodle production line is running, and establish a process node dataset; an anomaly sensor construction module, used to extract matching features from the process chain, call an anomaly identification library according to the matching features, and establish an anomaly sensor mapped to the process node using the anomaly identification library; an independent anomaly identification module, used to perform process identification of the corresponding process node dataset using the anomaly sensor, and establish an independent anomaly identification result; a collaborative anomaly identification module, used to extract the process node dataset into a process state vector, and then perform cross-process influence analysis of the process state vector using a state propagation function, and establish a collaborative anomaly identification result; and an anomaly early warning module, used to compensate the independent anomaly identification result with the collaborative anomaly identification result, and then report an anomaly early warning.

[0014] One or more technical solutions provided in this invention have at least the following beneficial effects: After reading the process chain of the instant noodle frying production line, a perception layer is configured at each process node to achieve a structured understanding and comprehensive perception deployment of the entire production process. When the instant noodle frying production line is running, the perception layer is synchronously activated to perceive the production data of the process nodes, dynamically collecting real-time operating data from each process node, establishing a process node dataset, and forming a raw dataset that can be used for anomaly identification. Matching features are extracted from the process chain, and an anomaly identification library is called based on these features. An anomaly perceiver is then established based on the anomaly identification library and mapped to the process nodes, realizing a feature-driven identification model matching mechanism. This ensures that different process nodes use the most suitable anomaly identification algorithm, improving the accuracy of anomaly identification. The anomaly perceiver is used to perform process identification on the corresponding process node dataset, conducting preliminary, node-level anomaly diagnosis for each process, establishing independent anomaly identification results, and forming the first level of anomaly identification information. After extracting the process node dataset into process state vectors, the cross-process impact analysis of the process state vectors is performed using the state propagation function. The state transmission relationship between processes is modeled, the downstream chain effect caused by upstream anomalies is identified, multi-node collaborative diagnosis is realized, and collaborative anomaly identification results are established. After compensating the independent anomaly identification results with the collaborative anomaly identification results, an anomaly warning is reported. By fusing independent and collaborative identification results, the accuracy and robustness of anomaly judgment are enhanced, and finally, reliable anomaly warning output is achieved.

[0015] In summary, this invention, by configuring a sensing layer at each process node of the fried instant noodle production line, comprehensively collects data and establishes independent anomaly identification results. At the same time, it considers the linkage relationship between processes to carry out collaborative anomaly identification, effectively improving the accuracy of fault identification and the real-time nature of early warning. It can not only detect anomalies at individual process nodes in a timely manner, but also accurately judge the chain reaction of faults in the collaborative process of multiple processes, greatly reducing the fault detection time, improving the overall operational stability of the fried instant noodle production line, and reducing the risk of production interruption and product quality damage caused by faults.

[0016] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the method for fault monitoring and remote early warning of a fried instant noodle production line provided in an embodiment of the present invention.

[0018] Figure 2 This is a flowchart illustrating the process of establishing collaborative anomaly identification results in the method for fault monitoring and remote early warning of a fried instant noodle production line provided in this embodiment of the invention.

[0019] Figure 3 This is a schematic diagram of the structure of the instant noodle production line fault monitoring and remote early warning system provided in an embodiment of the present invention.

[0020] Explanation of reference numerals in the attached figures: Perception configuration module 10, production data perception module 20, anomaly sensor construction module 30, independent anomaly identification module 40, collaborative anomaly identification module 50, and anomaly early warning module 60. Detailed Implementation

[0021] This invention provides a method and system for fault monitoring and remote early warning of a fried instant noodle production line. This solves the technical problems of low fault identification accuracy and delayed early warning response caused by isolated monitoring of each process and lack of inter-process linkage analysis in the prior art. It achieves the technical effect of improving fault identification accuracy and real-time early warning, and realizing collaborative fault monitoring between processes.

[0022] Example 1, as Figure 1 As shown in the figure, this invention provides a method for fault monitoring and remote early warning of a fried instant noodle production line, the method comprising: Step S100: After reading the process chain of the fried instant noodle production line, configure a perception layer at each process node.

[0023] Furthermore, the perception layer includes a basic sensing unit, a noise filtering unit, and a multimodal data fusion interface. The sensing parameters of the basic sensing unit include temperature, humidity, pressure, vibration, sound, visual image, flow rate, and current. The noise filtering unit is used for data preprocessing of the basic sensing unit, and the multimodal data fusion interface is used for aligning and fusing multiple types of sensing data.

[0024] Specifically, the process chain refers to the logical sequence of multiple consecutive processes in the production of fried instant noodles. Typical processes include: noodle cutting, steaming, frying, cooling, and packaging. A process node is a specific processing unit within the process chain, possessing relatively independent control logic and equipment resources. The perception layer is a data acquisition and preliminary processing module deployed at each process node, primarily used to acquire equipment status, environmental parameters, and operational process data.

[0025] First, the process configuration file or real-time status interface of the instant noodle frying production line is read to extract the process chain information of the entire production line, including the name, location, equipment type, and upstream and downstream process relationships of each process. Then, an independent sensing layer is deployed for each process node. This sensing layer includes a basic sensing unit, a noise filtering unit, and a multimodal data fusion interface. The basic sensing unit is used to collect various environmental and operational parameters related to the process, including temperature, humidity, pressure, vibration, sound, visual images, flow rate, and current. For example, in the frying process node, a temperature sensor monitors real-time oil temperature changes, a humidity sensor detects the water vapor content in the environment, a pressure sensor senses the stress on the equipment, a vibration sensor monitors the stability of equipment operation, a sound sensor collects equipment operating noise, a camera captures the material state during the frying process, a flow rate sensor monitors the oil flow speed, and a current sensor detects the motor operating current. The raw data collected by the basic sensing unit is input into the noise filtering unit to remove invalid data caused by environmental interference, equipment aging, and other factors. For example, a low-pass filtering algorithm is used to remove high-frequency electromagnetic noise, or wavelet transform is used to remove periodic interference signals. Subsequently, multiple types of sensing data will be synchronized in time and aligned semantically through a multimodal data fusion interface to form standardized and highly consistent fused data. For example, the Kalman filter method can be used to align multiple types of sensing data on the time axis, establish a unified data format and sampling rate, and form a complete process node dataset, providing high-quality data support for subsequent fault monitoring and early warning.

[0026] Step S200: After the instant noodle production line is running, the production data perception of the process node in the perception layer is activated synchronously to establish the process node dataset.

[0027] Specifically, once the instant noodle frying production line is operational, it automatically detects the production status and sends a synchronization activation signal, triggering the sensing layers deployed at each process node to enter working mode. Each sensing layer, based on its configured sensor type, synchronously collects production data for the corresponding process according to a preset sampling frequency and mode, including but not limited to physical quantities such as temperature, humidity, pressure, vibration, sound, visual images, flow rate, and current. All collected data, after noise processing and fusion alignment, is structured and stored according to metadata such as timestamps, process identifiers, and parameter types, ultimately forming multiple process node datasets covering all process nodes. These process node datasets will serve as input for subsequent anomaly sensors, enabling fault identification and risk analysis for both single and multi-point faults.

[0028] Step S300: Extract matching features from the process chain, call the anomaly identification library based on the matching features, and establish an anomaly sensor that maps to the process nodes using the anomaly identification library.

[0029] Specifically, matching features refer to the set of features extracted from the process chain that can identify process characteristics and key parameters, including process type, process parameter range, and correlation between preceding and following processes. The anomaly identification database is a collection of known abnormal events, typical failure cases, and normal state comparison samples from a fried instant noodle production line. Data types include structured sensor data, images, audio, and process logs, serving as the basic data source for training the anomaly sensor. The anomaly sensor is an anomaly detection model built based on the matching features and the anomaly identification database, targeting specific process nodes and capable of identifying specific abnormal situations at those nodes.

[0030] The acquired process chain is analyzed to extract matching features for each process node, including the process type identifier, key monitoring parameters, quality standards, and inter-process correlation characteristics. Using these extracted matching features as indexes, the most relevant historical anomaly data and control data for each process are extracted from the anomaly identification database. Multiple data subsets are constructed, each containing sensor data change patterns, anomaly characteristics, and corresponding diagnostic results, with each subset corresponding to a process node. Further, based on these data subsets, machine learning algorithms such as Support Vector Machines, Deep Neural Networks, or Random Forests are used to train anomaly detection models. A dedicated anomaly perceptron model is built for each process node, and a mapping relationship between the anomaly perceptron and the process node is established to ensure that each process node has a corresponding anomaly monitoring mechanism. After training, the anomaly perceptron model is deployed on the sensing layer or edge computing node of the corresponding process node to achieve anomaly identification for each process node's dataset.

[0031] Step S400: Use the anomaly sensor to identify the process of the corresponding process node dataset and establish an independent anomaly identification result.

[0032] Specifically, the anomaly sensor model established in step S300 is applied to the process node dataset of the corresponding process node to perform anomaly detection. Each process node's anomaly sensor operates independently, continuously receiving and analyzing the process node data stream from the perception layer, performing feature extraction and pattern recognition on the data, and calculating the similarity between the current data and known anomaly patterns or the degree of deviation from normal conditions. For different types of anomalies, the anomaly sensor outputs the anomaly type identification result and the corresponding confidence score. All detected anomaly events are recorded, including attributes such as anomaly type, timestamp, anomaly location identifier, severity, and confidence level, forming the anomaly identification result for each process node. This anomaly identification result is based only on the data analysis of a single process node and does not consider the correlation between processes; therefore, it is recorded as an independent anomaly identification result.

[0033] Step S500: After extracting the process node dataset into process state vectors, the cross-process influence analysis of the process state vectors is performed using the state propagation function to establish collaborative anomaly identification results.

[0034] Specifically, the process node dataset for each process node within a specific time window is transformed into a process state vector to represent the current operating state of that process, thus constructing a state vector field for the entire instant noodle production line. A state propagation function is then invoked to analyze the propagation paths of these vectors in space (process sequence) and time, identifying possible propagation directions and coupling relationships of anomalies. The state propagation function is a mathematical model describing how abnormal states propagate and influence the process chain. Built based on historical data and a physical model, it can predict the degree of impact and delay time of an abnormal state in one process on other processes. Finally, by combining these propagation paths, it is inferred which nodes' anomalies may be caused by the influence of other nodes, thereby generating system-level collaborative anomaly identification results for tracing the causes of anomalies and risk chains.

[0035] Step S600: After compensating the independent anomaly identification result with the collaborative anomaly identification result, an anomaly warning is issued.

[0036] Specifically, after obtaining independent and collaborative anomaly identification results, the two are logically fused to correct potential limitations and biases in independent anomaly detection. First, the anomaly sources and propagation paths identified in the collaborative anomaly identification results are examined. Then, the independent anomaly identification results are adjusted in several ways: for intermediate process nodes located on the anomaly propagation path, their impact from upstream anomalies is supplemented and labeled according to the propagation path; for terminal nodes with obvious independent anomaly performance but only actually affected, their root cause anomaly rating is lowered, and they are reclassified as secondary anomalies. This source-tracing and compensation mechanism can more comprehensively and accurately reflect the actual anomaly situation and risks of each process node. After compensation, a more accurate comprehensive anomaly identification result is formed, and an early warning report is generated, including anomaly type, time, impact chain, and location of occurrence, and pushed to the operation and maintenance terminal through the HMI interface or industrial communication interface.

[0037] Furthermore, step S300 includes: Step S310: Establish a process chain sequence based on the process chain, wherein the processes in the process chain sequence include cutting, steaming, shaping, frying, cooling, and packaging.

[0038] Step S320: After extracting feature tuples for each process in the process chain sequence, extract the global features of the process chain sequence.

[0039] Step S330: Use the feature tuple and the global feature as matching features to call the anomaly detection library to establish an anomaly perceiver.

[0040] Specifically, the process chain sequence refers to the ordered set of technological steps in a fried instant noodle production line, reflecting the flow-oriented structure of the production process. Each node process has temporal dependencies and technological coupling relationships. Based on the production line control logic or the deployment sequence of on-site equipment, key technological nodes in the fried instant noodle manufacturing process are identified and coded, constructing the following typical process chain sequence: noodle cutting - steaming - shaping - frying - cooling - packaging. This sequence serves as the structural basis for subsequent multi-node feature extraction, model mapping, and anomaly propagation analysis.

[0041] Feature tuples are sets of parameters describing the characteristics of a single process step, containing multi-dimensional information such as key process parameters, equipment characteristics, and quality indicators. Global features are a set of features used to describe the overall operational status and process synergy of the entire process chain sequence in the instant noodle production line, reflecting the mutual constraints, temporal relationships, and system-level constraints between processes. For each process step, various types of raw data are read from the perception layer, and typical feature tuple examples are extracted as shown in Table 1. Table 1—Examples of typical characteristic tuples for process operations Process Feature tuple examples section {Cut shaft speed, raw material feed rate, cutting surface frequency fluctuation value} Steaming {Steam temperature, humidity fluctuations, and pressure stability indicators} forming {Pressure roller speed, dough sheet formation rate, thickness uniformity, peak value of forming vibration} Deep-fried {Oil temperature distribution, image color difference, foam density, current characteristics} cool down {Cooling air velocity, dough surface temperature gradient, equipment vibration frequency, moisture content} Package {Packaging film tensile strength, sealing temperature, image positioning offset, barcode recognition accuracy} Based on the feature tuples, the entire process chain sequence is further analyzed, and global features are extracted across nodes, including: overall operating cycle time, average material transfer delay between processes, and synchronization deviation between processes. The overall operating cycle time refers to the unit output time after all processes are synchronized, which measures the entire production line's ability to complete the entire processing flow within a unit of time; cycle time fluctuations indicate multi-point coordination misalignment. The average material transfer delay between processes refers to the average interval between the output of one process and the next, reflecting the smoothness of logistics between processes; increased delays may indicate upstream accumulation or downstream blockage. The synchronization deviation between processes refers to the statistical value of the offset of the start or stop time of each process, describing whether there are abnormal start delays or advances in the process chain.

[0042] The extracted feature tuples and global features of each process are used as matching features. Historical records are retrieved from the anomaly detection database. These matching features are then used to perform similarity matching queries within the database. Algorithms such as cosine similarity and Euclidean distance are employed to calculate the similarity between the current process features and records in the database, retrieving historical datasets most similar to the process characteristics of each process in the current production line. Each record in these datasets is labeled with a corresponding anomaly type. After obtaining the historical anomaly dataset, an anomaly perceiver for specific processes is trained. An example of the training process is as follows: The historical dataset retrieved from the anomaly detection database is cleaned, standardized, and denoised to ensure data quality. Based on process characteristics and data features, a suitable machine learning model architecture is selected, such as support vector machines, random forests, deep neural networks, or ensemble models. The model is trained using the historical dataset. During training, the model parameters are continuously adjusted to optimize performance. Precision, recall, and F1 score metrics are used to evaluate the model's performance on the test dataset to ensure it meets application requirements. The trained anomaly perceiver model is deployed to the corresponding process nodes. Each process node has a dedicated anomaly perceiver capable of accurately identifying various anomalies that may occur in that specific process.

[0043] Furthermore, such as Figure 2 As shown, step S500 includes: Step S510: Treat each process node of the fried instant noodle production line as a local state field source point, and perform time evolution according to the process state vector to establish a dynamic state field.

[0044] Step S520: After constructing the state coupling path based on the dynamic state field, the linkage strength and state influence between processes are analyzed using the state propagation function to construct a third-order cooperative tensor.

[0045] Step S530: Use the third-order collaborative tensor to trace the source of collaborative anomalies and establish collaborative anomaly identification results.

[0046] Specifically, each process node in the instant noodle production line (including noodle cutting, steaming, forming, frying, cooling, and packaging) is considered as an independent local state field source point. The process state vector of each process node is used as the initial input, and a sliding time window is used to model its state evolution trend, generating a time-varying process state sequence. Subsequently, these process state sequences are numerically modeled using a time evolution function (such as an exponential decay function or a Markov state transition model), thereby constructing a time-dimensional dynamic state field around each process node. This dynamic state field is a temporal state mapping formed on the basis of time evolution, which can be understood as a state trajectory field formed by the change of process states over time. The dynamic state field reflects the evolution law and spatial diffusion capability of process state values ​​between process nodes, providing a basis for subsequent state coupling.

[0047] By analyzing the temporal similarity and propagation delay between various dynamic state fields, a set of state coupling paths is constructed. These state coupling paths refer to paths where states influence each other between processes, such as a linkage path like "frying anomaly - cooling process temperature anomaly". Then, these state coupling paths are modeled, and the state propagation function is used to quantitatively evaluate the state propagation strength between nodes. In the process of constructing the state propagation function, the following technical path is adopted: First, the state vector sequence of the local state field source point i and the state vector sequence of the target node j are obtained respectively. The time sequence is aligned by the sliding window method. Then, the Pearson correlation coefficient Corr of the two is calculated to measure the consistency of the trend of state change of the process node. The value range of the correlation coefficient Corr is [-1, 1]. The larger the value, the stronger the influence of the local state field source point i on the target node j. At the same time, mutual information or Granger causality test is introduced as a supplementary indicator of nonlinearity or causality. The state propagation function is constructed in the following form: P(i, j, Δt) = α*Corr(Si(t), Sj(t+Δt)) + β*MI(Si, Sj), where Si(t) represents the state vector of node i at time t; Sj(t+Δt) represents the state vector of node j after a delay of Δt; Sj represents the state vector of node j at time t; MI represents the mutual information; α and β are adjustable weight factors used to control the degree of integration of linear and nonlinear factors; Δt is the propagation delay time. Finally, the quantification results of the state propagation intensity between process nodes over time are organized into a third-order collaborative tensor, with the structure T(i, j, t), representing the strength of the state influence of process node i on process node j in time period t. This third-order tensor characterizes the coupling relationship among multiple processes and time slices in the production process, providing high-dimensional structural support for subsequent collaborative anomaly identification and anomaly tracing.

[0048] Based on the previously constructed third-order collaborative tensor, state coupling paths are analyzed using tensor analysis techniques. Preferably, CP decomposition or Tucker decomposition methods can be used to represent the third-order collaborative tensor as a product of several potential factors, thereby identifying the dominant propagation factors and key coupling relationships in the influence paths. First, a known abnormal process node Na is selected as the starting point for anomaly tracing. Linkage strength analysis is performed on all propagation paths in the third-order collaborative tensor targeting this node. A linkage strength threshold is set, and upstream process nodes whose T(i, Na, t) is greater than the threshold are selected as potential sources of anomaly propagation. A weighted sum is calculated based on the anomaly confidence, state change amplitude, and state propagation strength of these upstream process nodes to calculate a collaborative anomaly score. Aggregate analysis is performed on the nodes with the highest scores to form a collaborative anomaly node set. This collaborative anomaly node set is used to comprehensively reflect the multi-node coupling influence relationships that may jointly lead to the current target node's anomaly, ultimately forming the collaborative anomaly identification result.

[0049] Furthermore, step S520 includes: Step S521: Configure the propagation of the source of interest using the independent anomaly identification results and the state coupling path.

[0050] Step S522: Using the propagation concern source point as the search starting point, perform a backtracking fluctuation search of the state coupling path and establish a backtracking fluctuation search result.

[0051] Step S523: After reconstructing the search starting point based on the backtracking fluctuation search results, the state propagation function is used to analyze the linkage strength and state influence between processes, and a third-order collaborative tensor is constructed.

[0052] Specifically, the propagation focus points refer to the process nodes that the anomaly sensor determines are currently abnormal. These process nodes serve as initial focus points for state propagation and linkage analysis. Based on the independent anomaly identification results, process nodes in the process chain that are detected as abnormal by the anomaly sensor are marked as propagation focus points. Combining the previously constructed state coupling path graph, these abnormal process nodes are used as root nodes to prepare for state tracing analysis.

[0053] Starting from the source of propagation concern, a backtracking analysis is performed upstream along the state coupling path to analyze the volatility of the dynamic state field of each preceding process node within the relevant time period. Indicators such as the rate of change of the mean squared error of the state, frequency drift, and trend abrupt changes are calculated to determine whether there are any abnormal causes. If a preceding process node is not marked as an independent anomaly, but its state fluctuations are strong and the fluctuation time is earlier than the occurrence time of the current source of propagation concern anomaly, it is considered a new candidate source of propagation concern and recorded as a backtracking volatility search result.

[0054] The identified candidate propagation concern points are merged with the original set of propagation concern points to form a new set of collaborative propagation starting points, and state propagation modeling is re-executed. The state propagation function is used to calculate the degree of influence of each propagation concern point on the state of its downstream nodes at different time periods, and these influences are quantified into an intensity matrix. A third-order collaborative tensor is then constructed in a three-dimensional structure. This third-order collaborative tensor is used to characterize the state linkage strength between process nodes at different times in the entire backtracking-propagation network, providing a comprehensive data structure for collaborative anomaly identification and tracing.

[0055] Furthermore, step S500, in establishing the collaborative anomaly identification results, also includes: Step S540: Introduce an oil condition sensor into the sensing layer and establish an oil monitoring time series.

[0056] Step S550: Use the oil monitoring time series to fit the oil deterioration and establish a time series evolution anomaly.

[0057] Step S560: Construct a time-series risk fluctuation path through the time-series evolution anomaly and state coupling path.

[0058] Step S570: Compensate the collaborative anomaly identification result using the time-series risk fluctuation path.

[0059] Specifically, oil quality monitoring time series is a sequence formed by the change of oil quality status data collected by oil quality status sensors over time, used to analyze the dynamic changes in oil quality status. Time-series evolution anomalies are abnormal changes in oil quality status over time, discovered through oil quality deterioration fitting, such as abnormal increases in oil temperature or excessively rapid darkening of oil color. Time-series risk fluctuation paths are paths constructed by combining time-series evolution anomalies and status coupling paths, reflecting abnormal fluctuations in oil quality status and their propagation risks.

[0060] An oil condition sensor is integrated into the sensing layer of the frying process node in a fried instant noodle production line. This sensor can measure key oil condition parameters such as polar compound content, color index, acid value, temperature, and usage time. These key oil condition parameters are periodically sampled, and an oil monitoring time series is constructed in chronological order. This series describes the changing trends and deterioration process of the frying oil during use, providing foundational data for subsequent risk analysis.

[0061] Curve fitting methods are used to model oil deterioration based on the collected oil monitoring time series. For example, multinomial regression, exponential decay models, or time series prediction models based on long short-term memory networks are used to fit the trend of oil quality changes over time or cumulative processing batches. When the output of the prediction model deviates from the actual oil parameters by more than a set threshold (e.g., more than 5%), it is determined that there is a time series evolution anomaly. This time series evolution anomaly reflects the potential negative impact of frying oil deterioration on the overall operating status of the production line. For example, during continuous frying, the dielectric constant and polar component content of the oil are recorded every 10 minutes by oil condition sensors, and training data is constructed using a sliding window (10 steps). A long short-term memory network is used to predict the oil condition trend for the next 30 minutes. If the predicted trend shows that the polar component will exceed the standard within 20 minutes, an early warning of oil deterioration is issued; if the residual between the actual observed value and the predicted value continues to rise and exceeds the set threshold, it is also considered a time series evolution anomaly.

[0062] Combining the aforementioned dynamic state field with the state coupling path between process nodes, temporal evolution anomalies are used as risk triggers, propagating to downstream related processes. Starting with the frying process, its impact on subsequent processes such as cooling and packaging is assessed along the state coupling path, constructing a temporal risk fluctuation path. This temporal risk fluctuation path reflects the chain of cascading anomalies caused by the deterioration of frying oil.

[0063] The aforementioned time-series risk fluctuation paths are integrated with the results of collaborative anomaly identification. Based on collaborative tensor analysis, an adjustment weight for oil deterioration anomaly factors is introduced to compensate for nodes with low existing anomaly scores but significantly affected by oil time-series fluctuations, thereby improving the identification coverage and early warning for complex coupled anomalies.

[0064] Furthermore, step S600 includes: Step S610: Generate an abnormal status identifier for each process node based on the independent anomaly identification results.

[0065] Step S620: Use the collaborative anomaly identification results to perform source tracing compensation for the anomaly status identifier.

[0066] Step S630: Report an anomaly warning based on the anomaly status identifier after source tracing and compensation.

[0067] Specifically, after compensating for independent anomaly identification results using collaborative anomaly identification results, the early warning output process begins, mainly including the following steps: First, based on the independent anomaly identification results for each process node, an initial anomaly status identifier is generated. This identifier includes whether an anomaly exists at each node, the probability value of the anomaly, its type, and severity. Next, these initial anomaly status identifiers are compared and fused with the collaborative anomaly identification results. If a process node is not marked as anomaly in the independent anomaly identification results, but its upstream or downstream nodes are significantly anomaly in the collaborative anomaly path, there may be a missed detection, which is compensated for using the collaborative results. Furthermore, if a process node is judged as anomaly in the independent anomaly identification results, but the collaborative analysis determines that it has not been significantly affected, its anomaly level can be appropriately lowered to reduce false alarms. Finally, based on the fused anomaly status identifiers, the early warning level is determined and a remote anomaly early warning report is generated, which is pushed to the remote operation and maintenance center via the HMI interface or industrial communication interface.

[0068] Furthermore, after issuing an anomaly warning, it also includes: Step S640: Create an early warning verification window based on the aforementioned abnormal warning.

[0069] Step S650: Determine whether there is a production line maintenance response in the early warning verification window. If there is no production line maintenance response, then perform a shutdown process for the instant noodle production line.

[0070] Specifically, after an anomaly status is reported and an anomaly warning is pushed, an early warning verification window is automatically configured for the current process node and its upstream and downstream related nodes. This window is a time period used to monitor whether there is manual confirmation, system response, or maintenance intervention. Its length can be dynamically adjusted according to the process risk level, for example, set to 3 to 10 minutes. Within the verification window, the system continuously monitors for response events. Specifically, it listens for any of the following response signals as criteria for maintenance response: personnel confirmation of operation, such as maintenance personnel confirming that the operation has been processed via HMI or mobile terminal; production line control system response action, such as triggering a local process speed reduction, parameter adjustment, or oil change command; and perceived data returning to normal, such as the abnormal indicators returning to a safe range within the verification window period. If no maintenance response signal is detected before the end of the early warning verification window, it is considered that the anomaly has not been handled, and an automatic shutdown mechanism is triggered according to the configured strategy. This mechanism can be executed according to the following strategies: limiting the process segment (e.g., only shutting down the frying and cooling modules); shutting down the entire line to prevent the anomaly from spreading; and recording the anomaly event log, forming a tracking number, and sending it to the remote operation and maintenance center. For example, if severe deterioration of the oil is detected during the frying process, an anomaly warning is immediately generated, and a 5-minute warning verification window is simultaneously established in the upstream and downstream processes (cooking and cooling). If no response action is detected within the window period, such as manual confirmation, oil replacement, or recovery of relevant sensor data, the frying module will automatically trigger a shutdown command after the window period ends, and the production line control system will be linked to shut down step by step to ensure production line safety.

[0071] Furthermore, after issuing the anomaly warning, the method also includes: Step S710: Obtain the early warning response of the fried instant noodle production line, and package and record the early warning response and the abnormal early warning as abnormal historical data.

[0072] Step S720: Call the abnormal historical data to identify the sensitive features of the fried instant noodle production line, and use the sensitive features to monitor and manage the fried instant noodles.

[0073] Specifically, after an anomaly warning is issued, a corresponding response monitoring channel is established to collect warning response data, including but not limited to the following types: operator feedback records (such as HMI panel confirmations, mobile terminal receipts); production line control system response actions, process parameter change logs, and fault handling logs. The aforementioned warning responses are packaged with the original anomaly warning information (such as anomaly type, process node, warning time, warning intensity, and collaborative tensor number) to generate standardized anomaly historical data entries, which are then recorded in a local database or remote intelligent operation and maintenance platform, forming a closed-loop traceability mechanism.

[0074] Periodically or through data-triggered mechanisms, stored historical anomaly data is mined and analyzed. Statistical and machine learning models (such as principal component analysis and random forests) are used to identify sensitive characteristics of the production line or specific processes. These sensitive characteristics are key features or links prone to anomalies identified through the analysis of historical anomaly data. For example: certain types of equipment (such as frying temperature control valves) are more prone to failure under high temperature and humidity conditions; the failure probability of a certain type of oil increases sharply after running for more than X hours; excessive material accumulation between the cooking and cooling processes can easily trigger cooling anomalies, etc. These sensitive characteristics are dynamically updated in the perception layer configuration and anomaly sensor model parameters to achieve scenario-specific monitoring optimization and early perception strategy adjustment, thereby improving the robustness and foresight of fault monitoring in the instant noodle production line.

[0075] In summary, the method for fault monitoring and remote early warning of instant noodle production line provided by the embodiments of the present invention has the following beneficial effects: After reading the process chain of the instant noodle frying production line, a perception layer is configured at each process node to achieve a structured understanding and comprehensive perception deployment of the entire production process. When the instant noodle frying production line is running, the perception layer is synchronously activated to perceive the production data of the process nodes, dynamically collecting real-time operating data from each process node, establishing a process node dataset, and forming a raw dataset that can be used for anomaly identification. Matching features are extracted from the process chain, and an anomaly identification library is called based on these features. An anomaly perceiver is then established based on the anomaly identification library and mapped to the process nodes, realizing a feature-driven identification model matching mechanism. This ensures that different process nodes use the most suitable anomaly identification algorithm, improving the accuracy of anomaly identification. The anomaly perceiver is used to perform process identification on the corresponding process node dataset, conducting preliminary, node-level anomaly diagnosis for each process, establishing independent anomaly identification results, and forming the first level of anomaly identification information. After extracting the process node dataset into process state vectors, the cross-process impact analysis of the process state vectors is performed using the state propagation function. The state transmission relationship between processes is modeled, the downstream chain effect caused by upstream anomalies is identified, multi-node collaborative diagnosis is realized, and collaborative anomaly identification results are established. After compensating the independent anomaly identification results with the collaborative anomaly identification results, an anomaly warning is reported. By fusing independent and collaborative identification results, the accuracy and robustness of anomaly judgment are enhanced, and finally, reliable anomaly warning output is achieved.

[0076] Overall, the embodiments of the present invention, by configuring a sensing layer at each process node of the fried instant noodle production line, comprehensively collect data and establish independent anomaly identification results, while considering the linkage relationship between processes for collaborative anomaly identification, effectively improve the accuracy of fault identification and the real-time nature of early warning. It can not only detect anomalies at individual process nodes in a timely manner, but also accurately judge the chain reaction of faults in the collaborative process of multiple processes, significantly reduce the fault detection time, improve the overall operational stability of the fried instant noodle production line, and reduce the risk of production interruption and product quality damage caused by faults.

[0077] Example 2, as Figure 3 As shown, based on the same inventive concept as in Embodiment 1 above, this embodiment of the invention provides a fault monitoring and remote early warning system for a fried instant noodle production line, the system comprising: The perception configuration module 10 is used to configure a perception layer at each process node after reading the process chain of the fried instant noodle production line.

[0078] The production data sensing module 20 is used to synchronously activate the sensing layer to sense the production data of the process nodes after the instant noodle production line is running, and to establish a process node dataset.

[0079] The anomaly sensor construction module 30 is used to extract matching features from the process chain, call the anomaly recognition library according to the matching features, and establish an anomaly sensor mapped to the process node using the anomaly recognition library.

[0080] The independent anomaly identification module 40 is used to identify the corresponding process node dataset using the anomaly sensor and establish an independent anomaly identification result.

[0081] The collaborative anomaly identification module 50 is used to extract the process node dataset into process state vectors, and then use the state propagation function to perform cross-process influence analysis of the process state vectors to establish collaborative anomaly identification results.

[0082] The anomaly warning module 60 is used to compensate the independent anomaly recognition result with the collaborative anomaly recognition result and then report an anomaly warning.

[0083] Furthermore, the perception layer includes a basic sensing unit, a noise filtering unit, and a multimodal data fusion interface. The sensing parameters of the basic sensing unit include temperature, humidity, pressure, vibration, sound, visual image, flow rate, and current. The noise filtering unit is used for data preprocessing of the basic sensing unit, and the multimodal data fusion interface is used for aligning and fusing multiple types of sensing data.

[0084] Furthermore, in this embodiment of the invention, the anomaly sensor construction module 30 is also used to perform the following steps: A process chain sequence is established based on the process chain, wherein the processes in the process chain sequence include cutting, steaming, shaping, frying, cooling, and packaging; after extracting feature tuples for each process in the process chain sequence, global features of the process chain sequence are extracted; the feature tuples and global features are used as matching features to call an anomaly detection library to establish an anomaly sensor.

[0085] Furthermore, in this embodiment of the invention, the collaborative anomaly identification module 50 is also used to perform the following steps: Each process node in the instant noodle production line is regarded as a local state field source point. A dynamic state field is established by performing time evolution based on the process state vector. After constructing the state coupling path based on the dynamic state field, the linkage strength and state influence between processes are analyzed using the state propagation function to construct a third-order collaborative tensor. The third-order collaborative tensor is used to trace the source of collaborative anomalies and establish collaborative anomaly identification results.

[0086] Furthermore, in this embodiment of the invention, the collaborative anomaly identification module 50 is also used to perform the following steps: The independent anomaly identification results and the state coupling path are used to configure the propagation concern source point; with the propagation concern source point as the search starting point, the backtracking fluctuation search of the state coupling path is performed to establish the backtracking fluctuation search result; after reconstructing the search starting point according to the backtracking fluctuation search result, the state propagation function is used to analyze the linkage strength and state influence between processes, and a third-order collaborative tensor is constructed.

[0087] Furthermore, in this embodiment of the invention, the collaborative anomaly identification module 50 is also used to perform the following steps: An oil condition sensor is introduced into the perception layer to establish an oil monitoring time series; the oil condition time series is used to fit oil deterioration and establish a time-series evolution anomaly; a time-series risk fluctuation path is constructed through the time-series evolution anomaly and the state coupling path; and the time-series risk fluctuation path is used to compensate for the collaborative anomaly identification results.

[0088] Furthermore, in this embodiment of the invention, the anomaly warning module 60 is also used to perform the following steps: Anomaly status identifiers for each process node are generated based on the independent anomaly identification results; the source of the anomaly status identifiers is traced and compensated using the collaborative anomaly identification results; and an anomaly warning is issued based on the source-compensated anomaly status identifiers.

[0089] Furthermore, in this embodiment of the invention, the anomaly warning module 60 is also used to perform the following steps: An early warning verification window is created based on the aforementioned abnormal warning; it is then determined whether a production line maintenance response exists in the early warning verification window. If no production line maintenance response exists, the instant noodle production line is shut down.

[0090] Furthermore, the system described in this embodiment of the invention is also used to perform the following steps: The system acquires early warning responses from the fried instant noodle production line, packages and records these responses and abnormal warnings as abnormal historical data, calls upon the abnormal historical data to identify sensitive features of the fried instant noodle production line, and uses these sensitive features to monitor and manage the fried instant noodles.

[0091] Through the foregoing detailed description of the method for fault monitoring and remote early warning of the fried instant noodle production line, those skilled in the art can clearly understand that the fault monitoring and remote early warning system for the fried instant noodle production line in this embodiment, as corresponding to the system disclosed in Embodiment 2, has corresponding functional modules and beneficial effects as it corresponds to the method disclosed in Embodiment 1. For relevant details, please refer to the description in the method section.

[0092] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for fault monitoring and remote early warning in a fried instant noodle production line, characterized in that, The method includes: After reading the process chain of the fried instant noodle production line, a perception layer is configured at each process node; Once the instant noodle production line is running, the perception layer is activated simultaneously to perceive the production data of the process nodes and establish a process node dataset. Extract matching features from the process chain, call the anomaly identification library based on the matching features, and establish an anomaly sensor that maps to the process nodes using the anomaly identification library; The anomaly sensor is used to identify the process of the corresponding process node dataset, and an independent anomaly identification result is established. After extracting the process node dataset into process state vectors, the cross-process influence analysis of the process state vectors is performed using the state propagation function to establish collaborative anomaly identification results. After compensating the independent anomaly identification results with the collaborative anomaly identification results, an anomaly warning is issued.

2. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 1, characterized in that, After extracting the process node dataset into process state vectors, the cross-process influence analysis of the process state vectors is performed using a state propagation function to establish collaborative anomaly identification results, including: Each process node in the fried instant noodle production line is regarded as a local state field source point, and a dynamic state field is established by performing time evolution based on the process state vector. After constructing the state coupling path based on the dynamic state field, the state propagation function is used to analyze the linkage strength and state influence between processes, and a third-order collaborative tensor is constructed. The third-order collaborative tensor is used to trace the source of collaborative anomalies and establish collaborative anomaly identification results.

3. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 2, characterized in that, After constructing the state coupling path based on the dynamic state field, the state propagation function is used to analyze the linkage strength and state influence between processes, and a third-order cooperative tensor is constructed, including: The independent anomaly identification results and the state coupling path are used to configure the propagation of the source of interest. Starting from the source of the propagation concern, a backtracking fluctuation search of the state coupling path is performed to establish the backtracking fluctuation search results; After reconstructing the search starting point based on the backtracking fluctuation search results, the state propagation function is used to analyze the linkage strength and state influence between processes, and a third-order collaborative tensor is constructed.

4. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 1, characterized in that, The step of extracting matching features from the process chain, calling an anomaly detection library based on the matching features, and establishing an anomaly sensor mapped to the process nodes using the anomaly detection library includes: A process chain sequence is established based on the process chain, wherein the processes in the process chain sequence include cutting, steaming, shaping, frying, cooling, and packaging; After extracting feature tuples for each process in the process chain sequence, the global features of the process chain sequence are extracted. The feature tuples and global features are used as matching features to call the anomaly detection library to establish an anomaly perceiver.

5. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 1, characterized in that, The perception layer includes a basic sensing unit, a noise filtering unit, and a multimodal data fusion interface. The sensing parameters of the basic sensing unit include temperature, humidity, pressure, vibration, sound, visual image, flow rate, and current. The noise filtering unit is used for data preprocessing of the basic sensing unit, and the multimodal data fusion interface is used for aligning and fusing multiple types of sensing data.

6. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 1, characterized in that, The step of compensating the independent anomaly identification result with the collaborative anomaly identification result and then reporting an anomaly warning includes: Anomaly status identifiers for each process node are generated based on the independent anomaly identification results; The collaborative anomaly identification results are used for source tracing and compensation of anomaly status identifiers; An anomaly warning is issued based on the anomaly status identifier after source tracing and compensation.

7. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 1, characterized in that, After the abnormal warning is issued, the following is included: Obtain the early warning response from the fried instant noodle production line, and package and record the early warning response and the abnormal early warning as abnormal historical data; The abnormal historical data is used to identify sensitive features of the fried instant noodle production line, and these sensitive features are used for monitoring and management of the fried instant noodles.

8. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 1, characterized in that, The establishment of collaborative anomaly identification results also includes: Introduce oil condition sensors into the sensing layer to establish an oil monitoring time series; The oil product monitoring time series was used to fit oil product deterioration and establish time-series evolution anomalies. A temporal risk fluctuation path is constructed by the aforementioned temporal evolution anomaly and state coupling path; The time-series risk fluctuation path is used to compensate for the collaborative anomaly identification results.

9. The method for fault monitoring and remote early warning of a fried instant noodle production line as described in claim 1, characterized in that, After the abnormal warning is issued, the following is included: Create an alert verification window based on the aforementioned anomaly alert; Determine whether there is a production line maintenance response in the warning verification window. If there is no production line maintenance response, then execute the shutdown process for the instant noodle frying production line.

10. A fault monitoring and remote early warning system for a fried instant noodle production line, characterized in that, The system is used to execute the method for fault monitoring and remote early warning of the instant noodle production line according to any one of claims 1-9, including: The perception configuration module is used to configure a perception layer at each process node after reading the process chain of the fried instant noodle production line. The production data sensing module is used to synchronously activate the sensing layer to sense the production data of the process nodes after the instant noodle production line is running, and to establish a process node dataset. An anomaly sensor construction module is used to extract matching features from the process chain, call the anomaly identification library based on the matching features, and establish an anomaly sensor that maps to the process nodes using the anomaly identification library. An independent anomaly identification module is used to identify the corresponding process node dataset using the anomaly sensor and establish an independent anomaly identification result. The collaborative anomaly identification module is used to extract the process node dataset into process state vectors, and then use the state propagation function to perform cross-process influence analysis of the process state vectors to establish collaborative anomaly identification results. An anomaly warning module is used to compensate the independent anomaly recognition result with the collaborative anomaly recognition result and then report an anomaly warning.

Images