A method and device for dynamic tracking of the entire process of power material delivery

Abstract

This invention discloses a method and device for dynamic tracking of the entire process of power material performance. The method includes: collecting multi-dimensional location data of power material orders based on dynamic order vouchers; generating dynamic spatiotemporal weights for each location point based on the multi-dimensional location data of the power material orders to form a standardized trajectory of the power material orders; constructing a multi-dimensional high-order feature tensor based on the standardized trajectory of the power material orders and combined with multimodal order data; analyzing the deviation of each dimension in the multi-dimensional high-order feature tensor, providing the deviation tensor of each dimension, and determining the comprehensive compliance matching degree based on the preset dimension weights; generating a dynamic adaptive compliance range based on the type of power material and the compliance correction coefficient; judging the comprehensive compliance matching degree through the dynamic adaptive compliance range, providing an order processing plan, and completing the dynamic tracking of the entire process of power material performance, realizing "full-link visibility, intelligent acceptance, and quantitative risk control" of the performance process.

Description

Technical Field

[0001] This invention belongs to the technical field of power material management, specifically relating to a method and device for dynamic tracking of the entire process of power material performance. Background Technology

[0002] In the field of power material performance management, existing technical systems mainly rely on single data source monitoring or post-event auditing models, making it difficult to achieve efficient, dynamic, and intelligent control over the entire performance process. Current systems generally lack the ability to integrate and intelligently reason about multi-source heterogeneous data such as logistics trajectories, on-site images, and order information in real time. This makes it impossible to transform complex business data into interpretable decision-making basis, directly leading to problems such as delayed detection of performance anomalies, unclear responsibility traceability chains, and inefficient review processes.

[0003] Patent CN105701617A discloses a visualization method for full-process information management and control of the power material supply chain. Its features include the following steps: receiving allocation requests; integrating the ERP system with the data center; summarizing and analyzing power inventory resources and contract resources; tracking business operations; and controlling available resource information; using ABAP to periodically collect and summarize the entire business process and alarm data from the ERP system; and using Oracle tasks to collect data to the data center in an incremental or full manner; after receiving the data, the data center performs data cleaning; and generating charts on the client side. This solution provides intuitive analysis functions for business tracking and monitoring early warning of the entire power material supply chain information management and control nodes, facilitating comprehensive tracking and monitoring of the entire power material supply chain.

[0004] The existing technologies mentioned above only organize and analyze order information in the supply process, resulting in limited accuracy in identifying performance risks, inability to provide early warnings and grade assessments of potential risks, and low overall management efficiency.

[0005] How to integrate and analyze multimodal data in the process of fulfilling contracts for power supplies, and achieve "full-link visibility, intelligent acceptance, and quantitative risk control" in the process of fulfilling contracts, is a problem that needs to be solved. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method and apparatus for dynamic tracking of the entire process of power material fulfillment. The method includes: collecting multi-dimensional location data of power material orders based on dynamic order vouchers; generating dynamic spatiotemporal weights for each location point based on the multi-dimensional location data to form a standardized trajectory for the power material orders; constructing a multi-dimensional high-order feature tensor based on the standardized trajectory of the power material orders and multi-modal order data; analyzing the deviations of each dimension in the multi-dimensional high-order feature tensor, providing deviation tensors for each dimension, and determining the comprehensive compliance matching degree based on preset dimension weights; generating a dynamic adaptive compliance range based on the type of power material and compliance correction coefficients; judging the comprehensive compliance matching degree through the dynamic adaptive compliance range, providing an order processing plan, and completing the dynamic tracking of the entire power material fulfillment process. By fusing and analyzing multi-modal data during the power material fulfillment process, "full-link visibility, intelligent acceptance, and quantitative risk control" of the fulfillment process is achieved.

[0007] In a first aspect, the present invention provides a method for dynamic tracking of the entire process of power material delivery, specifically including the following steps: Based on dynamic order vouchers, collect multi-dimensional location data of power material orders; Based on the multi-dimensional location data of power material orders, dynamic spatiotemporal weights of each location point are generated to form a standardized trajectory of power material orders. Based on the standardized trajectory of power material orders, and combined with multimodal order data, a multidimensional high-order feature tensor is constructed. Analyze the deviation of each dimension in the multi-dimensional high-order feature tensor, give the deviation tensor of each dimension, and determine the comprehensive compliance matching degree by combining the preset dimension weights. A dynamic adaptive compliance range is generated based on the type of power materials and the compliance correction coefficient. By dynamically and adaptively adjusting the compliance scope, the system assesses the overall compliance matching degree, provides order processing solutions, and completes dynamic tracking of the entire process of fulfilling power material obligations.

[0008] Furthermore, dynamic order vouchers are determined through the following steps: Collect multimodal order data for power supply orders; Based on preset compliance verification rules, multi-dimensional verification is performed on multimodal order data, and verification results are provided. Based on the verification results, a dynamic order voucher is generated.

[0009] Furthermore, based on the multi-dimensional location data of power supply orders, dynamic spatiotemporal weights are generated for each location point, forming a standardized trajectory for power supply orders, specifically including: Based on the data sources of the multi-dimensional positioning data corresponding to each positioning point, the source reliability coefficient of each positioning point is given; The frequency attenuation factor of each positioning point is determined by the time interval between the current positioning point and the previous positioning point. Based on the optimal transportation route of power material orders, the path deviation of each location point is analyzed, and the path deviation penalty term for each location point is given. By integrating the source credibility coefficient, frequency attenuation factor, and path deviation penalty term, the dynamic spatiotemporal weight of each positioning point is determined. Based on the dynamic spatiotemporal weights of each location point, and combined with the sliding spatiotemporal window mechanism, the location points in the spatiotemporal window are fitted and denoised to generate a standardized trajectory for power material orders.

[0010] Furthermore, based on the dynamic spatiotemporal weights of each location point and combined with a sliding spatiotemporal window mechanism, the location points within the spatiotemporal window are fitted and denoised to generate a standardized trajectory for power material orders, specifically including: Obtain the positioning coordinates of each positioning point; By integrating the positioning coordinates and dynamic spatiotemporal weights, the product of the positioning coordinates and dynamic spatiotemporal weights of each positioning point is calculated to obtain the first positioning term; Sum the first positioning items corresponding to each positioning point in the spatiotemporal window to give the window positioning items; Sum the dynamic spatiotemporal weights of each location point in the spatiotemporal window to generate a window weight term; Calculate the ratio of the window positioning term to the window weight term to determine the window trajectory; By integrating the window trajectories of various time and space windows, a standardized trajectory for power material orders is formed.

[0011] Furthermore, the multimodal order data includes order time data groups, order image data groups, and order information data groups for power material orders; the multidimensional high-order feature tensors include high-order feature tensors in the time dimension, spatial dimension, image dimension, and order dimension. Based on the standardized trajectories of power supply orders and combined with multimodal order data, a multi-dimensional high-order feature tensor is constructed, specifically including: By mapping the order time data set to the output, a high-order feature tensor in the time dimension is formed; Based on the standardized trajectory of power material orders, the trajectory deviation is analyzed, and a high-order feature tensor with spatial dimension is generated. Based on the order image data set, the order images are vectorized to determine the high-order feature tensors of the image dimensions; The data in the order information data group is processed to generate a high-order feature tensor for the order dimension.

[0012] Furthermore, the biases of each dimension in the multi-dimensional high-order feature tensor are analyzed, and the bias tensors of each dimension are given. Combined with the preset dimension weights, the comprehensive compliance matching degree is determined, specifically including: Normalize the multi-dimensional high-order feature tensor to obtain the multi-dimensional standard feature tensor; Singular value decomposition is performed on the multi-dimensional standard feature tensor to extract the core features of each dimension and form a multi-dimensional core feature tensor. By combining the standard feature tensor, the deviation between the multi-dimensional core feature tensor and the theoretical feature tensor of the corresponding dimension is calculated to obtain the deviation tensor of each dimension. By adjusting the sparse constraint coefficients and the violation feature weight matrix, the deviation tensor of each dimension is given, and the core deviation tensor of each dimension is given. By combining the preset dimension weights and integrating the core deviation tensors of each dimension, a comprehensive compliance matching degree is obtained.

[0013] Furthermore, by combining preset dimension weights and integrating the core deviation tensors of each dimension, a comprehensive compliance matching degree is obtained, specifically including: Based on the core deviation tensors of each dimension, determine the core deviation norm and the maximum deviation norm of each dimension; Calculate the ratio of the core deviation norm to the maximum deviation norm in each dimension to determine the deviation sub-items in each dimension; Based on the dimension penalty coefficient of each dimension, the deviation sub-items of each dimension are corrected and the preset dimension weights are integrated to obtain the initial matching degree. Based on the preset matching degree output range, the initial matching degree is output mapped to give a comprehensive compliance matching degree.

[0014] Furthermore, based on the type of power equipment and the compliance correction coefficient, a dynamic adaptive compliance range is generated, specifically including: Based on the type of power equipment, determine the multi-level risk range corresponding to each type of power equipment; Determine the compliance correction factor through the power supply supplier; By combining compliance correction coefficients, the multi-level risk range is adjusted to generate a dynamic adaptive compliance range.

[0015] Furthermore, it also includes: Based on the entire process of each power material order, it is divided into multiple discrete event nodes. The behavior status of each discrete event node is analyzed to determine the behavioral normative characteristics and time fluctuation characteristics of each power material order. Based on the preset behavioral norm weights, time fluctuation weights, and compliance matching weights, the control score of each power material order is determined by integrating the behavioral norm characteristics, time fluctuation characteristics, and comprehensive compliance matching degree of each power material order. By analyzing and judging the control scores of each power material order, the detection and control plan for the corresponding power materials in each power material order is determined, and the hierarchical control of power material orders is completed.

[0016] Secondly, the present invention also provides a dynamic tracking device for the entire process of power material performance, employing the dynamic tracking method for the entire process of power material performance as described above, including: The data acquisition module is used to collect multi-dimensional location data of power material orders based on dynamic order vouchers; The trajectory determination module is used to generate dynamic spatiotemporal weights for each location point based on multi-dimensional location data of power material orders, thereby forming a standardized trajectory for power material orders. The tensor construction module is used to construct multi-dimensional high-order feature tensors based on the standardized trajectory of power material orders and combined with multimodal order data. The compliance calculation module is used to analyze the deviation of each dimension in the multi-dimensional high-order feature tensor, give the deviation tensor of each dimension, and determine the comprehensive compliance matching degree by combining the preset dimension weights. The scope determination module is used to generate a dynamic adaptive compliance scope based on the type of power materials and the compliance correction coefficient. The results output module is used to judge the overall compliance matching degree through dynamic adaptive compliance scope, provide order processing solutions, and complete the dynamic tracking of the entire process of power material performance.

[0017] The present invention provides a method and apparatus for dynamic tracking of the entire process of power material performance, which has at least the following beneficial effects: (1) By collecting multi-dimensional location data of power material orders, dynamic spatiotemporal weights of each location point are generated to determine the standardized trajectory of power material orders; then, a multi-dimensional high-order feature tensor is constructed to analyze the deviation of each dimension, and combined with the preset dimension weights, the comprehensive compliance matching degree is determined. Finally, the comprehensive compliance matching degree is judged by the dynamic adaptive compliance range, and an order processing plan is given to complete the dynamic tracking of the entire process of power material performance, so as to realize the "full-link visibility, intelligent acceptance, and quantitative risk control" of the performance process.

[0018] (2) By using dynamic spatiotemporal weights to smooth the centroid, the problems of trajectory jumps and floating deviations caused by tunnel signal loss, tall building obstruction, and electromagnetic interference can be effectively suppressed. At the same time, by generating continuous, smooth, and noise-free standardized trajectories, the real-time location, estimated arrival time, and compliance of the driving route are simultaneously marked, so as to realize the reliable management and control of the entire transportation trajectory process.

[0019] (3) By constructing a four-dimensional high-order feature tensor of time-space-image-order, a data foundation is provided for determining the comprehensive compliance matching degree and improving the accuracy and efficiency of acceptance.

[0020] (4) By constructing residual sparse constraint terms, the deviation tensors of each dimension are optimized and adjusted to filter out non-violation minor deviations such as transportation bumps, shooting angles, and normal route fine-tuning, and to give the core deviation tensors of each dimension. Attached Figure Description

[0021] Figure 1 A flowchart illustrating the dynamic tracking method for the entire process of power material performance provided in this embodiment of the invention; Figure 2 A flowchart illustrating the process of determining a detection and control scheme provided in an embodiment of the present invention; Figure 3 The structural block diagram of the dynamic tracking device for the entire process of power material performance provided in the embodiments of the present invention.

[0022] The modules are as follows: 201. Data acquisition module; 202. Trajectory determination module; 203. Tensor construction module; 204. Compliance calculation module; 205. Range determination module; 206. Result output module. Detailed Implementation

[0023] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0024] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0025] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0026] like Figure 1 As shown in the figure, this embodiment of the invention provides a method for dynamic tracking of the entire process of power material delivery, and the specific steps are as follows: S101: Based on dynamic order vouchers, collect multi-dimensional location data of power material orders.

[0027] The dynamic order voucher is an electronic delivery document with a unique blockchain traceability code and encrypted identification code. It contains full details of each power material in the power material order, batch number, agreed delivery time and location, carrier information, and dynamic confirmation fields from all three parties. The dynamic order voucher can also be simultaneously pushed to the supplier, carrier, and power grid materials department. This achieves fully paperless and traceable power material orders, serving as the sole core credential for subsequent power material order tracking, acceptance verification, and compliance assessment.

[0028] Based on the dynamic order voucher, multi-dimensional location data of the corresponding power material orders is collected. This multi-dimensional location data is the location data collected during the transportation of power material orders, including vehicle GPS location data, driver's mobile terminal Beidou location data, road checkpoint video capture location data, logistics platform location data, vehicle IoT device location data, etc. Various location data have differences in data timestamp, coordinate format, sampling frequency, and accuracy, which require data preprocessing.

[0029] Furthermore, dynamic order vouchers are determined through the following steps: Collect multimodal order data for power supply orders; Based on preset compliance verification rules, multi-dimensional verification is performed on multimodal order data, and verification results are provided. Based on the verification results, a dynamic order voucher is generated.

[0030] In one specific implementation, the multimodal order data includes a full range of data such as purchase order number, material model, quantity, technical parameters, contractual delivery cycle, warehouse receiving capacity, construction site demand nodes, split / consolidated delivery ratio, and supplier qualifications. A preset compliance verification engine is activated, and based on preset compliance verification rules, the initial delivery plan submitted by the supplier undergoes multi-dimensional verification, including whether the single batch delivery volume exceeds the warehouse limit, whether it matches the construction schedule, whether it complies with the contractual split delivery agreement, whether there is any illegal cross-category mixing, and whether it meets special requirements for material transportation (such as shockproof and moisture-proof transportation specifications for power equipment). Non-compliant plans are automatically rejected, and the optimal split delivery plan and dynamic order vouchers are generated.

[0031] S102: Based on the multi-dimensional location data of power material orders, generate dynamic spatiotemporal weights for each location point to form a standardized trajectory for power material orders.

[0032] Specifically, based on multi-dimensional location data of power supply orders, dynamic spatiotemporal weights are generated for each location point, forming a standardized trajectory for power supply orders, including: Based on the data sources of the multi-dimensional positioning data corresponding to each positioning point, the source reliability coefficient of each positioning point is given; The frequency attenuation factor of each positioning point is determined by the time interval between the current positioning point and the previous positioning point. Based on the optimal transportation route of power material orders, the path deviation of each location point is analyzed, and the path deviation penalty term for each location point is given. By integrating the source credibility coefficient, frequency attenuation factor, and path deviation penalty term, the dynamic spatiotemporal weight of each positioning point is determined. Based on the dynamic spatiotemporal weights of each location point, and combined with the sliding spatiotemporal window mechanism, the location points in the spatiotemporal window are fitted and denoised to generate a standardized trajectory for power material orders.

[0033] Dynamic spatiotemporal weights are specifically represented as follows:

[0034] Among them, W i Let α be the dynamic spatiotemporal weight of the location point i. i,src Let β be the source confidence coefficient for location point i. i,freq γ is the frequency attenuation factor for location point i. i,dev Here, λ is the path deviation penalty term for location point i, Δt is the time decay coefficient, and Δt is the time decay factor. i Let L be the time interval between the current location point i and the previous location point, k be the deviation penalty coefficient, and L be the distance between the current location point i and the previous location point. opt For the optimal transportation route of power supply orders, d(p) i ,L opt Let be the Euclidean distance between the location point i and the optimal transportation path.

[0035] For multi-dimensional positioning data, a spatiotemporal confidence weight model with three-dimensional coupling of source credibility, time decay, and path deviation is constructed to determine the dynamic spatiotemporal weight of any trajectory point. The source credibility coefficient is precisely assigned based on the authority of the positioning data collection source. For example, the source credibility coefficient for vehicle-mounted GPS positioning sources is 1.0, for road checkpoint / vehicle IoT positioning sources it is 0.9, for logistics platform / third-party carrier positioning sources it is 0.8, and for manually reported / mobile passive positioning sources it is 0.6. The higher the credibility of the data source, the larger the corresponding source credibility coefficient value.

[0036] The frequency attenuation factor of the positioning point is used to weaken invalid trajectory points with excessively long time intervals or discontinuous sampling. The time attenuation coefficient is set according to the actual scenario. In this example, combined with the power material transportation scenario, the time attenuation coefficient is set to 0.1. The longer the time interval between the current positioning point and the previous positioning point, the smaller the corresponding frequency attenuation factor, and the lower the corresponding dynamic spatiotemporal weight, thus eliminating noisy points with no signal in the field time.

[0037] The path deviation penalty is used to penalize abnormal trajectory points that deviate from the preset optimal transportation path. The greater the deviation distance, the smaller the value of the path deviation penalty, and the lower the corresponding dynamic spatiotemporal weight, so as to accurately identify abnormal driving behaviors such as detours and deviations.

[0038] Furthermore, based on the dynamic spatiotemporal weights of each location point and combined with a sliding spatiotemporal window mechanism, the location points within the spatiotemporal window are fitted and denoised to generate a standardized trajectory for power material orders, specifically including: Obtain the positioning coordinates of each positioning point; By integrating the positioning coordinates and dynamic spatiotemporal weights, the product of the positioning coordinates and dynamic spatiotemporal weights of each positioning point is calculated to obtain the first positioning term; Sum the first positioning items corresponding to each positioning point in the spatiotemporal window to give the window positioning items; Sum the dynamic spatiotemporal weights of each location point in the spatiotemporal window to generate a window weight term; Calculate the ratio of the window positioning term to the window weight term to determine the window trajectory; By integrating the window trajectories of various time and space windows, a standardized trajectory for power material orders is formed.

[0039] The window trajectory is specifically represented as follows:

[0040] Where Traj(t) is the window trajectory of the spatiotemporal window at time t, window(t) is the spatiotemporal window centered at time t, and W i Coord(p) represents the dynamic spatiotemporal weights of the location point i. i Let be the coordinates of point i, where i is the point number. i ·Coord(p i ) is the first positioning item. Positioning items for the window. This is the window weight item.

[0041] Centroid smoothing enhanced by dynamic spatiotemporal weights effectively suppresses trajectory jumps and hovering deviations caused by tunnel signal loss, tall building obstruction, and electromagnetic interference. Simultaneously, by generating continuous, smooth, and noise-free standardized trajectories, and synchronously labeling real-time location, estimated arrival time, and route compliance, reliable control over the entire transportation trajectory process is achieved.

[0042] S103: Based on the standardized trajectory of power material orders, and combined with multimodal order data, construct a multi-dimensional high-order feature tensor.

[0043] Specifically, the multimodal order data includes order time data groups, order image data groups, and order information data groups for power material orders; the multidimensional high-order feature tensors include time dimension high-order feature tensors, spatial dimension high-order feature tensors, image dimension high-order feature tensors, and order dimension high-order feature tensors. Based on the standardized trajectories of power supply orders and combined with multimodal order data, a multi-dimensional high-order feature tensor is constructed, specifically including: By mapping the order time data set to the output, a high-order feature tensor in the time dimension is formed; Based on the standardized trajectory of power material orders, the trajectory deviation is analyzed, and a high-order feature tensor with spatial dimension is generated. Based on the order image data set, the order images are vectorized to determine the high-order feature tensors of the image dimensions; The data in the order information data group is processed to generate a high-order feature tensor for the order dimension.

[0044] In one specific implementation, the order time data set includes data such as actual arrival time, transportation time, and node time deviation. This data set is mapped to the [0,1] interval to form a high-order feature tensor in the time dimension. Based on the standardized trajectory of the power material orders, data such as trajectory endpoint coordinates, warehouse unloading area coordinates, and route offset are extracted from the standardized trajectory. The trajectory deviation between the standardized trajectory and the preset trajectory is calculated using Euclidean distance, providing a high-order feature tensor in the spatial dimension. The order image data set includes image data such as material nameplates, electronic seals, packaging integrity, and vehicle appearance. The image data in the order image data set is identified using a machine vision model and converted into digital feature vectors to determine the high-order feature tensor in the image dimension. The order information data set includes order information such as electronic delivery notes, contract-specified material models, quantities, batch numbers, and certificate numbers. The order data in the order information data set is organized to form a high-order feature tensor in the order dimension.

[0045] By constructing a four-dimensional high-order feature tensor of time, space, image, and order, a data foundation is provided for subsequently determining the comprehensive compliance matching degree and improving the accuracy and efficiency of acceptance.

[0046] S104: Analyze the deviation of each dimension in the multi-dimensional high-order feature tensor, give the deviation tensor of each dimension, and determine the comprehensive compliance matching degree by combining the preset dimension weights.

[0047] Specifically, the deviations of each dimension in the multi-dimensional high-order feature tensor are analyzed, the deviation tensors of each dimension are given, and the comprehensive compliance matching degree is determined by combining the preset dimension weights. This includes: Normalize the multi-dimensional high-order feature tensor to obtain the multi-dimensional standard feature tensor; Singular value decomposition is performed on the multi-dimensional standard feature tensor to extract the core features of each dimension and form a multi-dimensional core feature tensor. By combining the standard feature tensor, the deviation between the multi-dimensional core feature tensor and the theoretical feature tensor of the corresponding dimension is calculated to obtain the deviation tensor of each dimension. By adjusting the sparse constraint coefficients and the violation feature weight matrix, the deviation tensor of each dimension is given, and the core deviation tensor of each dimension is given. By combining the preset dimension weights and integrating the core deviation tensors of each dimension, a comprehensive compliance matching degree is obtained.

[0048] In one specific implementation, the multi-dimensional high-order feature tensors acquired in real time are first normalized to eliminate differences in dimensions and numerical magnitudes, resulting in multi-dimensional standard feature tensors. Then, singular value decomposition (SVD) is used to extract salient features from the multi-dimensional standard feature tensors, eliminating noisy and redundant features, retaining core features, compressing the tensor dimensions, and determining the multi-dimensional core feature tensors. After obtaining the multi-dimensional core feature tensors, hierarchical orthogonal decomposition is performed on the four dimensions of time, space, image, and order. The deviation matrix between the core feature tensor and the theoretical feature tensor of each dimension is calculated separately to avoid misjudgments caused by cross-dimensional feature interference.

[0049] The bias tensor is specifically represented as:

[0050] Among them, D d,init Let T be the bias tensor of dimension d. d,real For the core feature tensor of dimension d, T d,base Let d be the theoretical feature tensor. The closer the deviation tensor value is to 0, the higher the matching degree of that dimension; a large deviation value indicates a risk of violation in the corresponding dimension. For the order and image dimensions, orthogonal projection operations are introduced to calculate the cosine value of the angle between the feature vectors, accurately identifying violations such as nameplate tampering, seal replacement, and batch discrepancies. It's understandable that the core feature tensor is obtained from real-time collected data, while the theoretical feature tensor is pre-constructed and represents theoretical data.

[0051] By constructing the residual sparse constraint term using sparse constraint coefficients and the weight matrix of violation features, it is specifically represented as follows:

[0052] Where min() is the minimum value function, D all Let W be the concatenated tensor of the bias tensors of each dimension, ||·||1 be the L1 regularization sparsity constraint, τ be the sparsity constraint coefficient, and W be the concatenation tensor of each dimension. vio This is the weight matrix for violation features. The Hadamard product is used. The sparsity constraint coefficients and the violation feature weight matrix are set according to the actual situation. The violation feature weight matrix assigns high weights to core violation items such as material consistency, seal authenticity, and spatiotemporal matching, while assigning low weights to common appearance defects. The Hadamard product is used to achieve weighted amplification of violation features and sparse filtering of normal deviations.

[0053] By constructing residual sparse constraint terms, the deviation tensors of each dimension are optimized and adjusted to filter out non-violation-related minor deviations such as transportation bumps, shooting angles, and minor adjustments to normal routes, and to give the core deviation tensors of each dimension.

[0054] Specifically, by combining preset dimension weights and integrating the core deviation tensors of each dimension, a comprehensive compliance matching degree is obtained, which includes: Based on the core deviation tensors of each dimension, determine the core deviation norm and the maximum deviation norm of each dimension; Calculate the ratio of the core deviation norm to the maximum deviation norm in each dimension to determine the deviation sub-items in each dimension; Based on the dimension penalty coefficient of each dimension, the deviation sub-items of each dimension are corrected and the preset dimension weights are integrated to obtain the initial matching degree. Based on the preset matching degree output range, the initial matching degree is output mapped to give a comprehensive compliance matching degree.

[0055] Overall compliance matching degree, specifically expressed as:

[0056] Among them, CCF represents the overall compliance matching score. max w is the maximum value of the preset matching degree output range. d Let ||·|| be the dimensional weight of dimension d. F Let η be the Frobenius norm, max() be the maximum value function, and η be the maximum value function. d Let D be the dimensionality penalty coefficient for dimension d. d Let d be the core bias tensor. As the initial matching degree, For the deviation sub-item of dimension d, ||D d || F Let d be the core deviation norm, max(||D) d || F ) represents the maximum deviation norm of dimension d.

[0057] In a specific example, the matching score output range is [0, 100], then the Score max=100. The dimensional penalty coefficient is assigned a value of 1.2 for the order and image dimensions, and a value of 1.0 for the time and space dimensions. A higher overall compliance matching score indicates a higher degree of compliance and acceptance, and a lower risk of non-compliance.

[0058] For the scenario of power material performance acceptance, a composite computation system is constructed, which combines four-dimensional tensor hierarchical orthogonal decomposition, residual feature sparse constraints, and violation feature weighted mapping. This solves the industry pain points of traditional acceptance algorithms, such as poor generalization, inability to accurately locate minor violations, and susceptibility to targeted fraud.

[0059] S105: Generate a dynamic adaptive compliance range based on the type of power materials and the compliance correction coefficient.

[0060] Specifically, a dynamic adaptive compliance range is generated based on the type of power equipment and the compliance correction coefficient, including: Based on the type of power equipment, determine the multi-level risk range corresponding to each type of power equipment; Determine the compliance correction factor through the power supply supplier; By combining compliance correction coefficients, the multi-level risk range is adjusted to generate a dynamic adaptive compliance range.

[0061] S106: By dynamically and adaptively adjusting the compliance scope, the overall compliance matching degree is judged, an order processing solution is provided, and dynamic tracking of the entire process of power material performance is completed.

[0062] In one specific implementation, the importance and security impact of various power materials are determined based on their type, and a multi-level risk range is established. For example, if the power material type is core critical material (transformers, high-voltage cables, switchgear, and other network security equipment), the corresponding level 1 risk range [93,100], level 2 risk range [82,93), and level 3 risk range [0,82] are determined. If the overall compliance matching degree of the core critical material is within the corresponding level 1 risk range, it is directly released; if the overall compliance matching degree of the core critical material is within the corresponding level 2 risk range, it undergoes manual review; if the overall compliance matching degree of the core critical material is within the corresponding level 3 risk range, it is forcibly blocked. If the power material type is conventional auxiliary material (hardware accessories, ordinary consumables, etc.), the corresponding level 1 risk range [88,100], level 2 risk range [75,88], and level 3 risk range [0,75] are determined. If the overall compliance of routine auxiliary materials falls within the corresponding Level 1 risk range, they will be released directly; if the overall compliance of routine auxiliary materials falls within the corresponding Level 2 risk range, they will be manually reviewed; if the overall compliance of routine auxiliary materials falls within the corresponding Level 3 risk range, they will be forcibly blocked.

[0063] By analyzing the historical compliance records of power supply suppliers, a compliance correction coefficient is determined. In a specific example, 0.95 ≤ compliance correction coefficient ξ ≤ 1.05. If the power supply supplier is a high-compliance whitelist supplier, then ξ = 1.05, raising the lower limit of the first-level and second-level risk ranges to improve the performance efficiency of high-quality suppliers; if the power supply supplier is a key controlled supplier, then ξ = 0.95, lowering the lower limit of the first-level and second-level risk ranges to tighten the risk range and strengthen control.

[0064] By combining material risk level anchoring and historical compliance correction, differentiated and rigid judgments are made on the risk of power material categories and the historical performance data of suppliers, thereby improving the efficiency of performance risk assessment.

[0065] The method for dynamic tracking of the entire process of power material delivery also includes: Based on the entire process of each power material order, it is divided into multiple discrete event nodes. The behavior status of each discrete event node is analyzed to determine the behavioral normative characteristics and time fluctuation characteristics of each power material order. Based on the preset behavioral norm weights, time fluctuation weights, and compliance matching weights, the control score of each power material order is determined by integrating the behavioral norm characteristics, time fluctuation characteristics, and comprehensive compliance matching degree of each power material order. By analyzing and judging the control scores of each power material order, the detection and control plan for the corresponding power materials in each power material order is determined, and the hierarchical control of power material orders is completed.

[0066] The control score for power supply orders is specifically expressed as follows:

[0067] Among them, S control Let BNF be the control score for power material orders, TFF be the behavioral normative characteristics of power material orders, CCF be the comprehensive compliance matching degree, w1 be the behavioral normative weight, w2 be the time fluctuation weight, w3 be the compliance matching weight, H(S) be the behavioral entropy of power material orders, and P(S) be the control score for power material orders. j S represents the behavior and state of discrete event node j. j The corresponding probability of the state occurring, ln() is a logarithmic function, H max This represents the maximum entropy of power supply order behavior, i.e., the upper limit of the entropy value of power supply orders under a completely non-compliant state. `min()` is the minimum value function. The average time interval for each discrete event node. denoted as the standard deviation of the time interval between each discrete event node.

[0068] In a specific example, based on the key considerations of power material management, weights are set for behavioral norms, time fluctuations, and compliance matching. For example, the behavioral norms weight w1 = 0.4, the time fluctuation weight w2 = 0.3, and the compliance matching weight w3 = 0.3. In other examples, these can be adjusted according to the actual situation, and there are no restrictions on this.

[0069] In one specific implementation, refer to Figure 2 The entire process of each power material order is broken down into six discrete event nodes: order confirmation, delivery plan reporting, in-transit transportation, arrival inspection, document submission, and warehousing and archiving. Each discrete event node is assigned one of three behavioral states: compliance, minor deviation, or serious violation, forming a finite-state discrete random event sequence. Behavioral normative characteristics and time fluctuation characteristics of each power material order are extracted. By analyzing the behavioral states and probabilities of each discrete event node in the power material order and calculating the corresponding behavioral entropy, the behavioral normative characteristics of the power material order are determined. Behavioral entropy measures the degree of disorder and violation in the performance behavior throughout the entire power material order process; the lower the entropy value, the more standardized the behavior. By analyzing the time intervals of each discrete event node in the power material order, the time jitter of the power material order is calculated, determining the time fluctuation characteristics. Time jitter measures the degree of fluctuation in the time intervals of each discrete event node in the entire performance process of the power material order, reflecting the accuracy of time control; the smaller the jitter value, the more stable the time performance. Time jitter can accurately identify time delays and rhythmic disorder scenarios, quantifying time compliance.

[0070] Based on the control scores of each power material order, multi-level differentiated control is implemented for power material fulfillment to achieve precise and refined management. When the control score S of a power material order... control When the score is ≥90, an inspection-free channel is opened for power material orders, implementing an inspection-free testing and control scheme. This indicates that the entire process of the power material order is compliant, with minimal time fluctuations, and tensor matching meets standards, with no violations or abnormalities. No physical re-inspection is required; the order can be directly processed for warehousing and archiving. Simultaneously, the supplier is added to a high-credit whitelist, enjoying simplified procedures and priority acceptance privileges for subsequent contract fulfillment. If the control score of the power material order is 90 > S... control A score of ≥70 indicates a testing and control plan for random inspections of power material orders. This means that the power material orders have minor behavioral deviations or slight time fluctuations, but no substantive violations. Random inspections will be conducted on 10%-30% of the power materials, including re-inspections of key technical parameters and physical appearance. Minor deviations will be rectified, and the orders will be archived after passing the re-inspection. If the control score of the power material order is 70 > S... controlThis refers to the testing and control plan for mandatory inspection of power material orders. It indicates that there are obvious violations, significant delays, or low tensor matching in power material orders. In this case, a full-scale physical re-inspection of all materials and anomaly tracing will be initiated. The supplier's current performance rights will be suspended. The process can only be restarted after rectification. At the same time, the supplier's credit score will be deducted and the supplier will be included in the key control list.

[0071] The entire process of generating and analyzing data from power material orders creates a full-process order fulfillment report, which is then pushed to suppliers, carriers, power grid warehousing, and material management departments. Online digital encrypted signatures are used for verification, ensuring a complete and unforgeable record, replacing traditional paper signatures and improving archiving efficiency. Simultaneously, blockchain technology generates a unique hash value for the order fulfillment report, which is uploaded to a distributed blockchain node storage system, ensuring the data is immutable and permanently stored.

[0072] This invention achieves dynamic and accurate calculation of trajectory weights by constructing a three-dimensional coupled spatiotemporal confidence weighting algorithm based on source credibility, time decay, and path deviation. Combined with a weighted centroid smoothing trajectory algorithm enhanced by dynamic spatiotemporal weights, it solves the problems of distortion and jumps in multi-source heterogeneous trajectories. Simultaneously, it constructs a four-dimensional tensor fusion verification algorithm based on time, space, image, and order to automate and accurately verify material acceptance. Finally, it integrates the behavioral normative features and time fluctuation features of the entire power material order process with comprehensive compliance matching to achieve accurate compliance assessment, realize quantitative hierarchical closed-loop management of the entire process, and improve the efficiency of power material performance management.

[0073] Reference Figure 3 This invention provides a dynamic tracking device for the entire process of power material delivery, comprising: Data acquisition module 201 is used to collect multi-dimensional location data of power material orders based on dynamic order vouchers; The trajectory determination module 202 is used to generate dynamic spatiotemporal weights for each location point based on the multi-dimensional location data of the power material order, thereby forming a standardized trajectory for the power material order. Tensor construction module 203 is used to construct multi-dimensional high-order feature tensors based on the standardized trajectory of power material orders and combined with multimodal order data; The compliance calculation module 204 is used to analyze the deviation of each dimension in the multi-dimensional high-order feature tensor, give the deviation tensor of each dimension, and determine the comprehensive compliance matching degree by combining the preset dimension weights. The scope determination module 205 is used to generate a dynamic adaptive compliance scope based on the type of power materials and the compliance correction coefficient. The output module 206 is used to judge the overall compliance matching degree through dynamic adaptive compliance scope, provide order processing solutions, and complete the dynamic tracking of the entire process of power material performance.

[0074] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0075] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method for dynamic tracking of the entire process of power material delivery, characterized in that, include: Based on dynamic order vouchers, collect multi-dimensional location data of power material orders; Based on the multi-dimensional location data of power material orders, dynamic spatiotemporal weights of each location point are generated to form a standardized trajectory of power material orders. Based on the standardized trajectory of power material orders, and combined with multimodal order data, a multidimensional high-order feature tensor is constructed. Analyze the deviation of each dimension in the multi-dimensional high-order feature tensor, give the deviation tensor of each dimension, and determine the comprehensive compliance matching degree by combining the preset dimension weights. A dynamic adaptive compliance range is generated based on the type of power materials and the compliance correction coefficient. By dynamically and adaptively adjusting the compliance scope, the system assesses the overall compliance matching degree, provides order processing solutions, and completes dynamic tracking of the entire process of fulfilling power material obligations.

2. The method for dynamic tracking of the entire process of power material performance as described in claim 1, characterized in that, Dynamic order vouchers are determined through the following steps: Collect multimodal order data for power supply orders; Based on preset compliance verification rules, multi-dimensional verification is performed on multimodal order data, and verification results are provided. Based on the verification results, a dynamic order voucher is generated.

3. The method for dynamic tracking of the entire process of power material performance as described in claim 1, characterized in that, Based on multi-dimensional location data of power supply orders, dynamic spatiotemporal weights are generated for each location point, forming a standardized trajectory for power supply orders, specifically including: Based on the data sources of the multi-dimensional positioning data corresponding to each positioning point, the source reliability coefficient of each positioning point is given; The frequency attenuation factor of each positioning point is determined by the time interval between the current positioning point and the previous positioning point. Based on the optimal transportation route of power material orders, the path deviation of each location point is analyzed, and the path deviation penalty term for each location point is given. By integrating the source credibility coefficient, frequency attenuation factor, and path deviation penalty term, the dynamic spatiotemporal weight of each positioning point is determined. Based on the dynamic spatiotemporal weights of each location point, and combined with the sliding spatiotemporal window mechanism, the location points in the spatiotemporal window are fitted and denoised to generate a standardized trajectory for power material orders.

4. The method for dynamic tracking of the entire process of power material performance as described in claim 3, characterized in that, Based on the dynamic spatiotemporal weights of each location point and combined with a sliding spatiotemporal window mechanism, the location points within the spatiotemporal window are fitted and denoised to generate a standardized trajectory for power material orders, specifically including: Obtain the positioning coordinates of each positioning point; By integrating the positioning coordinates and dynamic spatiotemporal weights, the product of the positioning coordinates and dynamic spatiotemporal weights of each positioning point is calculated to obtain the first positioning term; Sum the first positioning items corresponding to each positioning point in the spatiotemporal window to give the window positioning items; Sum the dynamic spatiotemporal weights of each location point in the spatiotemporal window to generate a window weight term; Calculate the ratio of the window positioning term to the window weight term to determine the window trajectory; By integrating the window trajectories of various time and space windows, a standardized trajectory for power material orders is formed.

5. The method for dynamic tracking of the entire process of power material performance as described in claim 1, characterized in that, Multimodal order data includes order time data groups, order image data groups, and order information data groups for power material orders; multidimensional high-order feature tensors include high-order feature tensors in the time dimension, spatial dimension, image dimension, and order dimension. Based on the standardized trajectories of power supply orders and combined with multimodal order data, a multi-dimensional high-order feature tensor is constructed, specifically including: By mapping the order time data set to the output, a high-order feature tensor in the time dimension is formed; Based on the standardized trajectory of power material orders, the trajectory deviation is analyzed, and a high-order feature tensor with spatial dimension is generated. Based on the order image data set, the order images are vectorized to determine the high-order feature tensors of the image dimensions; The data in the order information data group is processed to generate a high-order feature tensor for the order dimension.

6. The method for dynamic tracking of the entire process of power material delivery as described in claim 5, characterized in that, This analysis examines the biases of each dimension in a multi-dimensional high-order feature tensor, provides the bias tensor for each dimension, and, combined with predefined dimension weights, determines the overall compliance matching degree. Specifically, this includes: Normalize the multi-dimensional high-order feature tensor to obtain the multi-dimensional standard feature tensor; Singular value decomposition is performed on the multi-dimensional standard feature tensor to extract the core features of each dimension and form a multi-dimensional core feature tensor. By combining the standard feature tensor, the deviation between the multi-dimensional core feature tensor and the theoretical feature tensor of the corresponding dimension is calculated to obtain the deviation tensor of each dimension. By adjusting the sparse constraint coefficients and the violation feature weight matrix, the deviation tensor of each dimension is given, and the core deviation tensor of each dimension is given. By combining the preset dimension weights and integrating the core deviation tensors of each dimension, a comprehensive compliance matching degree is obtained.

7. The method for dynamic tracking of the entire process of power material performance as described in claim 6, characterized in that, By combining preset dimension weights and integrating the core deviation tensors of each dimension, a comprehensive compliance matching degree is obtained, which specifically includes: Based on the core deviation tensors of each dimension, determine the core deviation norm and the maximum deviation norm of each dimension; Calculate the ratio of the core deviation norm to the maximum deviation norm in each dimension to determine the deviation sub-items in each dimension; Based on the dimension penalty coefficient of each dimension, the deviation sub-items of each dimension are corrected and the preset dimension weights are integrated to obtain the initial matching degree. Based on the preset matching degree output range, the initial matching degree is output mapped to give a comprehensive compliance matching degree.

8. The method for dynamic tracking of the entire process of power material performance as described in claim 1, characterized in that, Based on the type of power equipment and the compliance correction factor, a dynamic adaptive compliance range is generated, specifically including: Based on the type of power equipment, determine the multi-level risk range corresponding to each type of power equipment; Determine the compliance correction factor through the power supply supplier; By combining compliance correction coefficients, the multi-level risk range is adjusted to generate a dynamic adaptive compliance range.

9. The method for dynamic tracking of the entire process of power material performance as described in claim 8, characterized in that, Also includes: Based on the entire process of each power material order, it is divided into multiple discrete event nodes. The behavior status of each discrete event node is analyzed to determine the behavioral normative characteristics and time fluctuation characteristics of each power material order. Based on the preset behavioral norm weights, time fluctuation weights, and compliance matching weights, the control score of each power material order is determined by integrating the behavioral norm characteristics, time fluctuation characteristics, and comprehensive compliance matching degree of each power material order. By analyzing and judging the control scores of each power material order, the detection and control plan for the corresponding power materials in each power material order is determined, and the hierarchical control of power material orders is completed.

10. A dynamic tracking device for the entire process of power material delivery, characterized in that, The method for dynamic tracking of the entire process of power material performance as described in any one of claims 1-9 includes: The data acquisition module is used to collect multi-dimensional location data of power material orders based on dynamic order vouchers; The trajectory determination module is used to generate dynamic spatiotemporal weights for each location point based on multi-dimensional location data of power material orders, thereby forming a standardized trajectory for power material orders. The tensor construction module is used to construct multi-dimensional high-order feature tensors based on the standardized trajectory of power material orders and combined with multimodal order data. The compliance calculation module is used to analyze the deviation of each dimension in the multi-dimensional high-order feature tensor, give the deviation tensor of each dimension, and determine the comprehensive compliance matching degree by combining the preset dimension weights. The scope determination module is used to generate a dynamic adaptive compliance scope based on the type of power materials and the compliance correction coefficient. The results output module is used to judge the overall compliance matching degree through dynamic adaptive compliance scope, provide order processing solutions, and complete the dynamic tracking of the entire process of power material performance.

Images