Food package fault monitoring method and system

Abstract

The invention discloses a food package fault monitoring method and system, and particularly relates to the field of food package fault monitoring, and the method comprises the steps: extracting a structure response variable of a food package object in an operation environment, constructing a disturbance expression sequence, and taking the disturbance expression sequence as a basis for path evolution recognition and industrial data processing analysis in the food package fault monitoring method; the method comprises the following steps: constructing a structure coupling mapping graph by analyzing a physical dependency relationship between food packaging structure arrangement and adjacent units; the change trend of a path disturbance parameter sequence formed by disturbance parameters corresponding to all paths in a structural path set is analyzed. By constructing a multi-level identification mechanism based on a structure disturbance expression sequence and a path propagation chain, industrial data capture and closed-loop traceability of a hidden defect propagation trend in a structure interdependence relationship between food packaging units are realized, so that the problem that the overall structure instability of a cluster type package cannot be identified in time is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of food packaging failure monitoring, and more particularly, to a food packaging failure monitoring method and system. Background Art

[0002] In the current food packaging distribution process, cluster packaging has become a common method to improve transportation efficiency and shelf management capabilities. However, existing inspection mechanisms often focus on the sealing, structural integrity, or appearance quality of individual packaging units, ignoring the structural interdependence between packaging units in the spatial layout.

[0003] As a result, if a unit has hidden defects, such as abnormal internal stress, micro-leakage, or insufficient sealing strength, even if there is no visible damage, it may cause local structural imbalance due to transportation vibration, temperature changes, or stacking compressive stress. This can trigger secondary batch risks such as deformation transfer to surrounding packaging units, content contamination, or stacking collapse, making the entire package that was originally qualified fail during actual circulation.

[0004] This type of problem, in which the hidden defects of a single food packaging unit cause the instability of the entire packaging structure, is difficult to be captured and traced in a timely manner in the industrial data processing process, so it will become the most easily overlooked weak link in the reliability assessment of food packaging. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a food packaging fault monitoring method and system. By constructing a multi-level identification mechanism based on structural perturbation expression sequences and path propagation chains, industrial data capture and closed-loop traceability of the hidden defect propagation trends in the structural interdependence between food packaging units are achieved, so as to solve the problem that the overall structural instability of clustered packaging cannot be identified in a timely manner.

[0006] To achieve the above objectives, the present invention provides the following technical solutions: a method for monitoring food packaging failures, comprising:

[0007] By extracting the structural response variables of food packaging objects under operating conditions, a disturbance expression sequence is constructed as the basis for path evolution identification and industrial data processing and analysis in food packaging fault monitoring methods.

[0008] By analyzing the physical dependencies between the food packaging structure and adjacent units, a structural coupling map is constructed.

[0009] By analyzing the changing trends of the path perturbation parameter sequence composed of the perturbation parameters corresponding to each path in the structural path set, a risk path chain with continuous evolution characteristics and the potential to cause cluster packaging structural instability is identified, which serves as the basis for food packaging failure monitoring, early warning, judgment and risk tracing.

[0010] By inverting the historical disturbance expression process of high-risk paths, the irreversible trend of disturbance evolution is identified, and based on this, it is determined whether the food packaging has entered the early warning stage;

[0011] Idle cycles are used to perform interpolation detection on the target area in the early warning path combination set and update the score map to realize industrial data processing feedback of the failure and collapse risk of food packaging.

[0012] In a preferred embodiment, after the food packaging is completed, the structural response variables of the food packaging object are collected within a specified time period. The structural response variables include three types of monitoring indicators: internal pressure change, surface capacitance change, and local temperature gradient change.

[0013] The structural response variables are segmented into equal width segments according to the time sliding window mechanism, and the peak filtering, boundary smoothing and median normalization operations are performed on the structural response variables of equal width segments to obtain the standardized response data sequence.

[0014] In a preferred embodiment, a disturbance feature is extracted for each data segment in the standardized response data sequence;

[0015] Determine whether the disturbance feature quantities extracted from any three adjacent time window segments in the standardized response data sequence are consistent in their changing direction. The disturbance feature quantities include the change slope value, the oscillation amplitude value, and the transient disturbance ratio value. If at least two of the three show a trend of changing in the same direction within the three segments, then the three segments are defined as potential continuous disturbance segments and are included in the abnormal trend cache queue as structural disturbance anomaly input;

[0016] All potential continuous perturbation segments are combined in time order and expression vector compression is performed to obtain a set of perturbation expression sequences;

[0017] The perturbed expression sequence set is input into the subsequent structural pathway construction process as an input benchmark for initial industrial data processing.

[0018] In a preferred embodiment, based on the food packaging sample groups to which the perturbed expression sequences of the food packaging objects belong, the spatial combination of the food packaging sample groups during stacking and boxing is used to extract the packaging arrangement rules and the boxing matrix structure, and based on this, an adjacency pair set is generated. The adjacency pair set includes three categories: direct physical contact pairs, pressure coupling pairs, and boundary deformation connection pairs.

[0019] Determine whether the adjacent pairs in the adjacent pair set satisfy both the pressure coupling relationship and the boundary deformation response continuity. If so, mark them as valid coupling relationships.

[0020] In a preferred embodiment, for each effective coupling relationship, the peak value of the amplitude difference is extracted from the disturbance expression sequence of the corresponding two packaging objects and compared with the structural fatigue threshold. If it is greater than the threshold, an initial disturbance conduction channel is established;

[0021] Determine whether there is a continuous two-stage propagation package object non-perturbation expression sequence response in the initial perturbation conduction channel. If so, disconnect the channel and mark it as a non-continuous path segment;

[0022] All effective conduction channels are integrated into a structural path set according to the spatial path direction. Each path in the path set is marked with directionality, disturbance amplitude gradient and cross-level propagation factor. The output structural path set serves as the structural baseline map for identifying high-risk chains in food packaging failure monitoring in the next stage.

[0023] In a preferred embodiment, the path length, inter-node disturbance amplitude, and number of node jumps are extracted for each path in the structural path set to form a path disturbance parameter group; it is determined whether there are three consecutive nodes in the path disturbance parameter group with an increasing disturbance amplitude, and if so, it is marked as a trend-increasing path;

[0024] Perform disturbance gradient accumulation calculation on the trend rising path and perform deviation matching with the stable disturbance baseline of the normal packaging object to obtain the path evolution offset value;

[0025] If the path evolution offset value is greater than the preset path evolution offset reference threshold, it is added to the risk disturbance path candidate set.

[0026] In a preferred embodiment, a multivariate cluster analysis is performed on the risk disturbance path candidate set. The multivariate cluster analysis takes the path disturbance parameter group in the risk disturbance path candidate set as input. The path disturbance parameter group includes three types of clustering parameters: gradient change rate, path connectivity index, and propagation direction consistency coefficient. The path group with a feature expression concentration greater than a preset concentration threshold is extracted from the clustering results based on the parameter distribution density as the initial selection of the risk disturbance path set.

[0027] The path group with a feature expression concentration greater than the preset concentration threshold extracted from the multivariate cluster analysis results is confirmed as the final risk disturbance path set, and this risk disturbance path set is used as the input path set for performing historical disturbance evolution trend inversion and reversibility judgment in the subsequent fault evolution backtracking process.

[0028] In a preferred embodiment, each path in the risk disturbance path set is traced back to the three period data segments before the activation period to extract the inversion input disturbance sequence; it is determined whether the disturbance change direction in the inversion input disturbance sequence remains consistent within the three periods, and if so, it is marked as a monotonic improvement path;

[0029] The difference between the historical disturbance fluctuation range of the monotonic improvement path and the fluctuation reference range of the standard stable path is evaluated. If the difference is within the preset variation range, the path is defined as a "reversible path", otherwise it is determined to be an "irreversible path";

[0030] If the path is determined to be reversible, the warning triggering operation for the current path is skipped, and the path disturbance information is returned to the path evolution trend update process for subsequent trend evaluation; if the path is determined to be irreversible, the starting node position corresponding to the path is extracted and marked as the potential crash starting point;

[0031] A cross-path superposition judgment is performed on all potential collapse starting points. If a node appears in two or more irreversible paths at the same time, it constitutes a warning fusion node; all paths containing warning fusion nodes are output, and a warning path combination set is constructed as the target area for interpolation detection in the next stage.

[0032] In a preferred embodiment, in the early warning path combination set, the disturbance expression sequence of the next cycle is extracted for each path according to the activation cycle to form an interpolation detection sample group. It is determined whether the disturbance amplitude of the corresponding path segment in the interpolation detection sample group continues to increase in two consecutive cycles. If so, the path is marked as an aggravated segment.

[0033] Perform disturbance residual re-estimation on all aggravated segment paths and superimpose them with the path gradient change of the previous cycle to form a superimposed disturbance amplitude value; determine whether the superimposed disturbance amplitude value is higher than twice the mean value of the global disturbance residual of the structural path; if so, mark the path segment as a key intervention segment;

[0034] A stability score calculation is performed on the key intervention section path. The scoring indicators include three types of variables: path length, cumulative disturbance and continuous aggravation coefficient; it is determined whether the stability score is lower than the dynamic risk threshold. If so, it is fed back to the structural adjustment strategy process.

[0035] A food packaging fault monitoring system includes a disturbance sequence module, a map-building module, a chain-judgment and risk identification module, a tracing and alarm determination module, and a supplementary measurement closed-loop module;

[0036] The disturbance sequence extraction module extracts the structural response variables of the food packaging object under the operating environment and constructs a disturbance expression sequence, which serves as the basis for path evolution identification and industrial data processing and analysis in the food packaging fault monitoring method;

[0037] The mapping and chaining module constructs a structural coupling map by analyzing the physical dependencies between the food packaging structure and adjacent units.

[0038] The chain-based risk identification module analyzes the changing trends of the path disturbance parameter sequence composed of the disturbance parameters corresponding to each path in the structural path set, and identifies the risk path chain with continuous evolution characteristics that may cause cluster packaging structure instability. This serves as the basis for food packaging failure monitoring, early warning, judgment, and risk tracing.

[0039] The retrospective warning module inverts the historical disturbance expression process of high-risk paths, identifies the irreversible trend of disturbance evolution, and judges whether to enter the early warning stage of food packaging based on this;

[0040] The supplementary test closed-loop module uses idle cycles to perform interpolation detection on the target area in the early warning path combination set and update the score map, realizing industrial data processing feedback of the failure and collapse risk of food packaging.

[0041] The technical effects and advantages of the present invention are as follows:

[0042] 1. This method constructs a multi-path evolution identification process based on the structural disturbance propagation chain to achieve early prediction and path location of the risk of structural instability of the entire packaging group caused by a single hidden defect in industrial data processing, thus filling the blind spot of the existing assessment mechanism in identifying clustered risks.

[0043] 2. By using the consistency judgment of disturbance direction to construct a disturbance expression sequence, it is possible to capture the trend of weak structural anomalies on the time scale, laying a foundation for highly recognizable industrial data expression for subsequent structural path mapping and data modeling;

[0044] 3. Generate a set of structural paths through physical coupling relationship mapping, introduce multi-dimensional constraints such as pressure coupling and boundary deformation, achieve spatial modeling of the internal coupling mechanism of cluster packaging and screen out abnormal channels, and improve the robustness of the monitoring model under actual stacking conditions;

[0045] 4. Utilize the trend changes and clustering mechanism of the path perturbation parameter group to identify high-risk path chains, and then determine the irreversibility of evolution through perturbation inversion to achieve closed-loop tracing of potential structural collapse paths, ensuring that the monitoring system has temporal foresight and evolutionary explanatory power;

[0046] 5. Combined with idle cycles to perform interpolation detection and build a multi-factor stable scoring function, intervention judgment and scoring feedback are performed on high-risk path segments, building a fault monitoring system with continuous evaluation capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flowchart of the framework of the method steps of the present invention.

[0048] Figure 2 A flow chart was constructed for the perturbed expressed sequences of the present invention.

[0049] Figure 3A flow chart is constructed for the structural path coupling of the present invention.

[0050] Figure 4 This is a high-risk path identification flow chart of the present invention.

[0051] Figure 5 This is a flow chart of disturbance inversion and reversibility judgment of the present invention.

[0052] Figure 6 This is a flow chart of interpolation detection and scoring feedback of the present invention.

[0053] Figure 7 Schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] Refer to the instruction manual Figure 1-7 A food packaging fault monitoring method according to an embodiment of the present invention includes:

[0056] By extracting the structural response variables of food packaging objects under operating conditions, a disturbance expression sequence is constructed as the basis for path evolution identification and industrial data processing and analysis in food packaging fault monitoring methods.

[0057] By analyzing the physical dependencies between the food packaging structure and adjacent units, a structural coupling map is constructed.

[0058] By analyzing the changing trends of the path perturbation parameter sequence composed of the perturbation parameters corresponding to each path in the structural path set, we identify the risk path chain with continuous evolution characteristics that may cause cluster packaging structure instability, which serves as the basis for food packaging failure monitoring, early warning, judgment and risk tracing.

[0059] By inverting the historical disturbance expression process of high-risk paths, the irreversible trend of disturbance evolution is identified, and based on this, it is determined whether the food packaging has entered the early warning stage;

[0060] Idle cycles are used to perform interpolation detection on the target area in the early warning path combination set and update the score map to realize industrial data processing feedback of the failure and collapse risk of food packaging.

[0061] After the food packaging is completed, the structural response variables of the food packaging object are collected within a specified time period. The structural response variables include three types of monitoring indicators: internal pressure change, surface capacitance change, and local temperature gradient change;

[0062] The structural response variables are segmented into equal width segments according to the time sliding window mechanism, and the peak filtering, boundary smoothing and median normalization operations are performed on the structural response variables of equal width segments to obtain the standardized response data sequence.

[0063] Extracting disturbance feature quantity for each data segment in the standardized response data sequence;

[0064] Determine whether the disturbance feature quantities extracted from any three adjacent time window segments in the standardized response data sequence are consistent in their changing direction. The disturbance feature quantities include the change slope value, the oscillation amplitude value, and the transient disturbance ratio value. If at least two of the three show a trend of changing in the same direction within the three segments, then the three segments are defined as potential continuous disturbance segments and are included in the abnormal trend cache queue as structural disturbance anomaly input;

[0065] All potential continuous perturbation segments are combined in time order and expression vector compression is performed to obtain a set of perturbation expression sequences;

[0066] The perturbed expression sequence set is input into the subsequent structural pathway construction process as an input benchmark for initial industrial data processing.

[0067] Based on the food packaging sample groups to which the perturbation expression sequences of the food packaging objects belong, the spatial combination of the food packaging sample groups during the stacking and packing process is used to extract the packaging arrangement rules and packing matrix structure. Based on this, an adjacency pair set is generated. The adjacency pair set includes three categories: direct physical contact pairs, pressure coupling pairs, and boundary deformation connection pairs.

[0068] Determine whether the adjacent pairs in the adjacent pair set satisfy both the pressure coupling relationship and the boundary deformation response continuity. If so, mark them as valid coupling relationships.

[0069] For each effective coupling relationship, the peak value of the amplitude difference is extracted from the disturbance expression sequence of the corresponding two packaged objects and compared with the structural fatigue threshold. If it is greater than the threshold, the initial disturbance conduction channel is established.

[0070] Determine whether there is a continuous two-stage propagation package object non-perturbation expression sequence response in the initial perturbation conduction channel. If so, disconnect the channel and mark it as a non-continuous path segment;

[0071] All effective conduction channels are integrated into a structural path set according to the spatial path direction. Each path in the path set is marked with directionality, disturbance amplitude gradient and cross-level propagation factor. The output structural path set serves as the structural baseline map for identifying high-risk chains in food packaging failure monitoring in the next stage.

[0072] For each path in the structural path set, the path length, inter-node disturbance amplitude, and number of node jumps are extracted to form a path disturbance parameter group. It is determined whether there are three consecutive nodes in the path disturbance parameter group with an increasing disturbance amplitude. If so, it is marked as a trend-increasing path.

[0073] Perform disturbance gradient accumulation calculation on the trend rising path and perform deviation matching with the stable disturbance baseline of the normal packaging object to obtain the path evolution offset value;

[0074] If the path evolution offset value is greater than the preset path evolution offset reference threshold, it is added to the risk disturbance path candidate set.

[0075] Performing a multivariate cluster analysis on the risk perturbation path candidate set, the multivariate cluster analysis takes the path perturbation parameter group in the risk perturbation path candidate set as input, the path perturbation parameter group includes three types of clustering parameters: gradient change rate, path connectivity index, and propagation direction consistency coefficient. Based on the parameter distribution density, the path group with a feature expression concentration greater than a preset concentration threshold is extracted from the clustering results as the initial selection of the risk perturbation path set;

[0076] The path group with a feature expression concentration greater than the preset concentration threshold extracted from the multivariate cluster analysis results is confirmed as the final risk disturbance path set, and this risk disturbance path set is used as the input path set for performing historical disturbance evolution trend inversion and reversibility judgment in the subsequent fault evolution backtracking process.

[0077] Each path in the risk perturbation path set is traced back to the three period data segments before the activation period to extract the inversion input perturbation sequence; determine whether the perturbation change direction in the inversion input perturbation sequence remains consistent within the three periods. If so, mark it as a monotonic improvement path;

[0078] The difference between the historical disturbance fluctuation range of the monotonic improvement path and the fluctuation reference range of the standard stable path is evaluated. If the difference is within the preset variation range, the path is defined as a "reversible path", otherwise it is determined to be an "irreversible path";

[0079] If the path is determined to be reversible, the warning triggering operation for the current path is skipped, and the path disturbance information is returned to the path evolution trend update process for subsequent trend evaluation; if the path is determined to be irreversible, the starting node position corresponding to the path is extracted and marked as the potential crash starting point;

[0080] A cross-path superposition judgment is performed on all potential collapse starting points. If a node appears in two or more irreversible paths at the same time, it constitutes a warning fusion node; all paths containing warning fusion nodes are output, and a warning path combination set is constructed as the target area for interpolation detection in the next stage.

[0081] In the early warning path combination set, the disturbance expression sequence of the next cycle is extracted for each path according to the activation cycle to form an interpolation detection sample group. It is determined whether the disturbance amplitude of the corresponding path segment in the interpolation detection sample group continues to rise in two consecutive cycles. If so, it is marked as an aggravated segment path;

[0082] The disturbance residuals of all aggravated path segments are re-estimated and superimposed with the gradient change of the path in the previous cycle to form a superimposed disturbance amplitude value. The superimposed disturbance amplitude value is determined to be higher than twice the mean of the global disturbance residual of the structural path. If so, the path segment is marked as a key intervention segment. The selection of "twice" as the judgment criterion is based on the common empirical rule of anomaly detection in industrial data processing: when a disturbance value exceeds twice the mean of the overall residual, it usually indicates that its degree of deviation has exceeded the normal fluctuation range and is statistically significant. It is used to effectively identify potential high-risk path segments and intervene in structural stability in advance.

[0083] A stability score calculation is performed on the key intervention segment path. The scoring indicators include three types of variables: path length, cumulative disturbance and continuous aggravation coefficient. It is determined whether the stability score is lower than the dynamic risk threshold. If so, it is fed back to the structural adjustment strategy process to implement recommendations for eliminating batches of food packaging failures or optimizing the packing structure.

[0084] A food packaging fault monitoring system includes a disturbance sequence module, a map-building module, a chain-judgment and risk identification module, a tracing and alarm determination module, and a supplementary measurement closed-loop module;

[0085] The disturbance sequence extraction module extracts the structural response variables of the food packaging object under the operating environment and constructs a disturbance expression sequence, which serves as the basis for path evolution identification and industrial data processing and analysis in the food packaging fault monitoring method;

[0086] The mapping and chaining module constructs a structural coupling map by analyzing the physical dependencies between the food packaging structure and adjacent units.

[0087] The chain-based risk identification module analyzes the changing trends of the path disturbance parameter sequence composed of the disturbance parameters corresponding to each path in the structural path set, identifying risk path chains with continuous evolution characteristics that may cause cluster packaging structure instability. This serves as the basis for food packaging failure monitoring, early warning, judgment, and risk tracing.

[0088] The retrospective warning module inverts the historical disturbance expression process of high-risk paths, identifies the irreversible trend of disturbance evolution, and judges whether to enter the early warning stage of food packaging based on this;

[0089] The supplementary test closed-loop module uses idle cycles to perform interpolation detection on the target area in the early warning path combination set and update the score map, realizing industrial data processing feedback of the failure and collapse risk of food packaging.

[0090] It should be noted that in the formula structure involved in this solution, dimensionless terms can serve as proportionality or structural adjustment factors. When combined with quantities with units, they only play a numerical scaling role and do not introduce new physical dimensions. Therefore, they will not change or confuse the overall unit system of expression. This combination of "dimensionless terms and units" can be understood as a composite structural expression commonly used in mathematical and physical modeling, conforming to the principle of dimensional consistency and having a clear physical interpretation basis.

[0091] Secondly, in the formula structure of this scheme, if multiple variables with different physical units are involved, including but not limited to time, mass or energy variables, their joint appearance is to express the collaborative modeling relationship of multiple physical mechanisms. Each variable can be formed into a unified structure through function mapping, ratio combination or normalization adjustment. The units and meanings are clear, and the overall expression conforms to the principle of dimensional consistency and the common formula of engineering modeling.

[0092] In this solution, any design constants, weights, adjustment factors, threshold parameters, and proportional coefficients are adjustable control parameters for different application environments. Their values ​​depend on the target device configuration, data input characteristics, and performance optimization goals. During the implementation phase, they are set within a reasonable range through model verification, performance constraints, or engineering calibration. Although these parameters do not have unique preset values, they have clear adjustment logic and calculation paths, and are part of the deterministic setting process in engineering implementation. The purpose of such setting is to ensure that the solution is both universally adaptable, reproducible, and operable, without affecting its technical clarity and feasibility.

[0093] In constructing the perturbation expression sequence, time periods with consistent continuous perturbation directions are extracted from the food packaging structural response to generate a set of perturbation expression sequences as the industrial data basis for path construction.

[0094] The set of perturbation expression sequences is described by the formula:

[0095]

[0096] The perturbation consistency function Ψ(t) is defined as:

[0097] Ψ(t)=δ s (t)+δ a (t)+δr (t)

[0098] The consistency of each type of perturbation direction is defined as:

[0099]

[0100] where Δx t ∈{Δs t ,Δa t ,Δr t}

[0101] Where ε is the final set of perturbation expression sequences, which is used as the input for path mapping construction; time window t represents the time window index of the current response data segment; T is the total number of segments in the structure response data sequence; S t is the response segment of the normalized structural response sequence in the time window t; Δs t is the change in the slope of the structural response in the t-th time window; Δa t is the change in the structural response amplitude in the t-th time window; Δr t is the change value of the structural response disturbance ratio in the t-th time window; δ s (t),δ a (t),δ r (t) represents the Boolean function value indicating whether the directions of the three types of disturbances are consistent in the three time windows of t, t+1, and t+2; sign(·) is the sign function, which is used to determine the direction of change of the disturbance variable; is a gate function that determines whether a condition is met. If so, it returns 1; otherwise, it returns 0. In addition, T-3 indicates the last starting segment index of up to three consecutive time windows that can be formed starting from the 0th time window.

[0102] In identifying high-risk disturbance path chains, a disturbance parameter group is extracted from the structural path set, a path risk trend function is constructed, and multivariate clustering is performed to identify high-risk path chains;

[0103] Risk trend score by path p To express the path risk trend function:

[0104]

[0105] Where the cumulative disturbance amplitude Λ on path p p Expressed as:

[0106]

[0107] The maximum jump value Θ of the perturbation amplitude on path p p for:

[0108]

[0109] Consistency score Γ of continuous perturbation directions on path p p Expressed as:

[0110]

[0111] Cluster center extraction formula (targeting the highest risk trend concentration):

[0112]

[0113] Where p is the number of any structure path in the structure path set; n p is the number of disturbance nodes contained in path p; ΔF i is the disturbance amplitude difference of the i-th disturbance node on path p, ΔF i-1 is the disturbance amplitude difference of the i-1th disturbance node on path p; is the set of paths in cluster k; The high-risk path clusters finally extracted are used as the input of the warning path set;

[0114] In interpolation detection and dynamic stability scoring, real-time disturbance tracking of the warning path is performed through interpolation samples, and residual trends are superimposed and calculated to construct a multi-factor dynamic scoring function to mark key intervention path segments;

[0115] The dynamic risk score D of the jth path segment in the early warning path combination set j , to express the multi-factor dynamic scoring function:

[0116]

[0117] The path segment disturbance residual accumulation Σ j Expressed as:

[0118]

[0119] The path segment disturbance amplitude is a monotonically increasing trend count Φ within the interpolation period. j :

[0120]

[0121] Among them L j is the length of path segment j (in number of time windows); ΔR t is the change in the disturbance amplitude of the path at time t, ΔR t+1 is the change in the disturbance amplitude of the path at time t+1; τ jis the interpolation monitoring start period of path segment j; ξ, η, ζ are adjustment parameters representing path length, disturbance residual, and trend contribution factor, respectively (not specific values, which can be obtained through interpolation estimation or optimization in actual applications). The values ​​of ξ, η, ζ include but are not limited to manual assignment based on actual engineering experience or historical data, for example, setting a higher weight when a certain item has a greater impact, such as ξ:η:ζ = 2:1:1; the values ​​of ξ, η, ζ also include but are not limited to data fitting optimization, using existing data with labeled risk levels to perform parameter fitting by minimizing the scoring error, such as the least squares method, multi-objective optimization, etc.

[0122] This paper proposes a food packaging fault monitoring method based on an in-depth analysis of the structural chain instability that clustered packaging is prone to in actual distribution environments. Traditional packaging quality inspection methods focus on visible defects in individual packaging units, such as seal integrity or surface damage, but ignore the fact that under complex environments such as transportation vibration, temperature fluctuations, or stacking stress, minor structural anomalies in individual packages may cause structural collapse of the entire group of packages through physical contact, internal stress conduction, or deformation extension, thereby affecting the stability and safety of the entire batch of products. Therefore, based on the evolutionary logic of the "structural propagation chain", this solution designs a complete fault monitoring method driven by disturbance data and centered on path evolution identification, achieving early identification and process tracing of overall failures caused by local defects.

[0123] After encapsulation, the system collects structural response variables, including internal pressure, surface capacitance, and local temperature gradients, from the packaged food object. This data is then normalized using a time sliding window mechanism to extract three key disturbance features: slope of change, oscillation amplitude, and disturbance ratio. This step involves more than simple data observation. Instead, the system determines whether these features maintain consistency in their direction of change across multiple consecutive time windows. This allows the system to construct time periods with potential abnormal trends, known as disturbance expression sequences. These sequences not only condense key structural variation features but also provide input for the subsequent establishment of spatial propagation pathways.

[0124] The method then constructs a structural path mapping graph by mapping the spatial relationships of food packaging objects during stacking and packing. This graph structure is not a static topology, but rather integrates three types of adjacency relationships: physical contact, pressure coupling, and boundary continuity. Perturbation transmission channels are established based on the amplitude difference between perturbation expression sequences and the structural fatigue threshold. If a channel experiences a response interruption or path discontinuity along the propagation path, the path is removed in real time, ensuring that the constructed structural path set has high transmission reliability and spatial coupling integrity, thereby establishing a spatial carrier for dynamic fault identification for the entire method.

[0125] Based on this set of structural paths, the solution further analyzes the distribution trends of disturbance parameters along the paths, constructing a set of disturbance parameters including path length, disturbance amplitude, and jump amplitude. It then performs trend increase determination and calculates evolutionary offset values. These parameters reflect not only the spatial diffusion trend of the disturbance along the paths but also the temporal evolution of the disturbance intensity. A multivariate clustering algorithm is then used to further screen out highly concentrated disturbance links, resulting in a set of high-risk path chains that are highly likely to cause clustered structural instability. This process essentially continuously compresses the possible fault space and improves the targeted nature of fault identification.

[0126] To verify whether the evolution of these risky paths is inherently irreversible, a disturbance path inversion mechanism is introduced. Each high-risk path is traced back to its pre-activation historical period to reconstruct the disturbance change process. The disturbance fluctuation range is then compared with the disturbance reference value range of a normal, stable path to determine whether it exhibits a monotonically increasing and irreversible evolutionary trend. If an irreversible path is determined, the path starting position is extracted and a path cross-fusion judgment is performed to screen out key node groups that may be fault sources. This is then output as a set of warning paths to be interpolated and tested.

[0127] Finally, during idle time periods, periodic interpolation detection is performed on the warning path, continuously tracking the residual fluctuations of the disturbance. A dynamic stability scoring function is constructed, which uses path length, total disturbance amount, and continuous aggravation trend as scoring criteria to determine whether the current path segment is approaching the structural instability boundary. If the score result falls below the dynamic threshold, the system will trigger structural adjustment recommendations, such as local packaging structure optimization or batch elimination, ultimately realizing a closed-loop monitoring and industrial data processing feedback mechanism for food packaging failures.

[0128] In general, this method starts from structural response, takes disturbance expression construction as the basis, path coupling as the bridge, trend identification as the core, inversion judgment as the verification mechanism, and scoring closed loop as the control means to construct a full-process method for intelligent food packaging fault monitoring, providing a solution path for the safety assurance of large-scale cluster packaging in actual industrial applications.

[0129] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for monitoring food packaging failure, comprising: By extracting the structural response variables of food packaging objects under operating conditions, a disturbance expression sequence is constructed as the basis for path evolution identification and industrial data processing and analysis in food packaging fault monitoring methods. Its characteristics are: By analyzing the physical dependencies between the food packaging structure and adjacent units, a structural coupling map is constructed. By analyzing the changing trends of the path perturbation parameter sequence composed of the perturbation parameters corresponding to each path in the structural path set, a risk path chain with continuous evolution characteristics and the potential to cause cluster packaging structural instability is identified, which serves as the basis for food packaging failure monitoring, early warning, judgment and risk tracing. By inverting the historical disturbance expression process of high-risk paths, the irreversible trend of disturbance evolution is identified, and based on this, it is determined whether the food packaging has entered the early warning stage; Idle cycles are used to perform interpolation detection on the target area in the early warning path combination set and update the score map to realize industrial data processing feedback of the failure and collapse risk of food packaging.

2. A food packaging failure monitoring method according to claim 1, characterized in that: After the food packaging is completed, the structural response variables of the food packaging object are collected within a specified time period. The structural response variables include three types of monitoring indicators: internal pressure change, surface capacitance change, and local temperature gradient change; The structural response variables are segmented into equal width segments according to the time sliding window mechanism, and the peak filtering, boundary smoothing and median normalization operations are performed on the structural response variables of equal width segments to obtain the standardized response data sequence.

3. A food packaging failure monitoring method according to claim 2, characterized in that: Extracting disturbance feature quantity for each data segment in the standardized response data sequence; Determine whether the disturbance feature quantities extracted from any three adjacent time window segments in the standardized response data sequence are consistent in their changing direction. The disturbance feature quantities include the change slope value, the oscillation amplitude value, and the transient disturbance ratio value. If at least two of the three show a trend of changing in the same direction within the three segments, then the three segments are defined as potential continuous disturbance segments and are included in the abnormal trend cache queue as structural disturbance anomaly input; All potential continuous perturbation segments are combined in time order and expression vector compression is performed to obtain a set of perturbation expression sequences; The perturbed expression sequence set is input into the subsequent structural pathway construction process as an input benchmark for initial industrial data processing.

4. A food packaging failure monitoring method according to claim 3, characterized in that: Based on the food packaging sample groups to which the perturbation expression sequences of the food packaging objects belong, the spatial combination of the food packaging sample groups during the stacking and packing process is used to extract the packaging arrangement rules and packing matrix structure. Based on this, an adjacency pair set is generated. The adjacency pair set includes three categories: direct physical contact pairs, pressure coupling pairs, and boundary deformation connection pairs. Determine whether the adjacent pairs in the adjacent pair set satisfy both the pressure coupling relationship and the boundary deformation response continuity. If so, mark them as valid coupling relationships.

5. A food packaging failure monitoring method according to claim 4, characterized in that: For each effective coupling relationship, the peak value of the amplitude difference is extracted from the disturbance expression sequence of the corresponding two packaged objects and compared with the structural fatigue threshold. If it is greater than the threshold, the initial disturbance conduction channel is established. Determine whether there is a continuous two-stage propagation package object non-perturbation expression sequence response in the initial perturbation conduction channel. If so, disconnect the channel and mark it as a non-continuous path segment; All effective conduction channels are integrated into a structural path set according to the spatial path direction. Each path in the path set is marked with directionality, disturbance amplitude gradient and cross-level propagation factor. The output structural path set serves as the structural baseline map for identifying high-risk chains in food packaging failure monitoring in the next stage.

6. A food packaging failure monitoring method according to claim 5, characterized in that: For each path in the structural path set, the path length, inter-node disturbance amplitude, and number of node jumps are extracted to form a path disturbance parameter group. It is determined whether there are three consecutive nodes in the path disturbance parameter group with an increasing disturbance amplitude. If so, it is marked as a trend-increasing path. Perform disturbance gradient accumulation calculation on the trend rising path and perform deviation matching with the stable disturbance baseline of the normal packaging object to obtain the path evolution offset value; If the path evolution offset value is greater than the preset path evolution offset reference threshold, it is added to the risk disturbance path candidate set.

7. A food packaging failure monitoring method according to claim 6, characterized in that: Performing a multivariate cluster analysis on the risk perturbation path candidate set, the multivariate cluster analysis takes the path perturbation parameter group in the risk perturbation path candidate set as input, the path perturbation parameter group includes three types of clustering parameters: gradient change rate, path connectivity index, and propagation direction consistency coefficient. Based on the parameter distribution density, the path group with a feature expression concentration greater than a preset concentration threshold is extracted from the clustering results as the initial selection of the risk perturbation path set; The path group with a feature expression concentration greater than the preset concentration threshold extracted from the multivariate cluster analysis results is confirmed as the final risk disturbance path set, and this risk disturbance path set is used as the input path set for performing historical disturbance evolution trend inversion and reversibility judgment in the subsequent fault evolution backtracking process.

8. A food packaging failure monitoring method according to claim 7, characterized in that: Each path in the risk perturbation path set is traced back to the three period data segments before the activation period to extract the inversion input perturbation sequence; determine whether the perturbation change direction in the inversion input perturbation sequence remains consistent within the three periods. If so, mark it as a monotonic improvement path; The difference between the historical disturbance fluctuation range of the monotonic improvement path and the fluctuation reference range of the standard stable path is evaluated. If the difference is within the preset variation range, the path is defined as "reversible path", otherwise it is determined to be "irreversible path"; If the path is determined to be reversible, the warning triggering operation for the current path is skipped, and the path disturbance information is returned to the path evolution trend update process for subsequent trend evaluation; if the path is determined to be irreversible, the starting node position corresponding to the path is extracted and marked as the potential crash starting point; Perform cross-path superposition judgment on all potential collapse starting points. If a node appears in two or more irreversible paths at the same time, it constitutes an early warning fusion node; Output all paths containing warning fusion nodes and construct a warning path combination set as the target area for interpolation detection in the next stage.

9. A food packaging failure monitoring method according to claim 8, characterized in that: In the early warning path combination set, the disturbance expression sequence of the next cycle is extracted for each path according to the activation cycle to form an interpolation detection sample group. It is determined whether the disturbance amplitude of the corresponding path segment in the interpolation detection sample group continues to rise in two consecutive cycles. If so, it is marked as an aggravated segment path; Perform disturbance residual re-estimation on all aggravated segment paths and superimpose them with the path gradient change of the previous cycle to form a superimposed disturbance amplitude value; determine whether the superimposed disturbance amplitude value is higher than twice the mean value of the global disturbance residual of the structural path; if so, mark the path segment as a key intervention segment; A stability score calculation is performed on the key intervention section path. The scoring indicators include three types of variables: path length, cumulative disturbance and continuous aggravation coefficient; it is determined whether the stability score is lower than the dynamic risk threshold. If so, it is fed back to the structural adjustment strategy process.

10. A food packaging fault monitoring system, comprising a disturbance sequence module, a map chain module, a chain judgment and risk identification module, a tracing and alarm determination module, and a supplementary measurement closed-loop module, characterized in that: The disturbance sequence extraction module extracts the structural response variables of the food packaging object under the operating environment and constructs a disturbance expression sequence, which serves as the basis for path evolution identification and industrial data processing and analysis in the food packaging fault monitoring method; The mapping and chaining module constructs a structural coupling map by analyzing the physical dependencies between the food packaging structure and adjacent units. The chain-based risk identification module analyzes the changing trends of the path disturbance parameter sequence composed of the disturbance parameters corresponding to each path in the structural path set, and identifies the risk path chain with continuous evolution characteristics that may cause cluster packaging structure instability. This serves as the basis for food packaging failure monitoring, early warning, judgment, and risk tracing. The retrospective warning module inverts the historical disturbance expression process of high-risk paths, identifies the irreversible trend of disturbance evolution, and judges whether to enter the early warning stage of food packaging based on this; The supplementary test closed-loop module uses idle cycles to perform interpolation detection on the target area in the early warning path combination set and update the score map, realizing industrial data processing feedback of the failure and collapse risk of food packaging.

Images