Supply chain risk real-time early warning and coping system

Abstract

The invention discloses a supply chain risk real-time early warning and response system, and relates to the technical field of supply chain early warning. A data acquisition module obtains a main supplier equipment state, a historical interruption record and standby supplier logistics route data in real time; the dynamic weight distribution module generates an alternative priority evaluation model based on the basic weight parameters and the sudden environment parameters; the substitution link pre-calculation module generates a multi-level substitution path according to the evaluation model and calculates a score; the risk quantification module calculates a main supplier risk value; when the risk value exceeds the threshold value, the instruction generation module calls the highest score to replace the path and sends a work order change instruction; the dynamic feedback module collects execution data, establishes a deviation relation and corrects an evaluation model weight rule; all the modules achieve a complete link from risk identification to coping strategy optimization through data interaction and calculation, and the real-time performance and accuracy of supply chain risk management are improved.

Description

Technical Field

[0001] The present invention relates to the field of supply chain early warning technology, and in particular to a real-time early warning and response system for supply chain risks. Background Art

[0002] In a globalized supply chain, companies face multiple risks, including unexpected supplier shutdowns, logistics network disruptions, and raw material price fluctuations. Traditional supply chain risk management techniques typically employ static assessment models based on historical data, passively defending against these risks by regularly checking supplier credit ratings or setting inventory safety thresholds. However, these techniques have significant drawbacks when addressing sudden and contagious risks: First, existing early warning systems rely on manually entered discrete data and are unable to capture dynamic information such as equipment operating status and logistics route changes in real time, causing risk identification to lag behind actual business processes. Second, the mechanism for selecting alternative suppliers often uses a fixed-weight scorecard, which fails to consider the dynamic impact of environmental variables such as sudden weather changes and policy adjustments on logistics timeliness and quality stability, resulting in an inadequate match between emergency plans and actual needs. Third, traditional solutions lack closed-loop control of execution feedback, and historical decision-making data fails to effectively feed back into the risk assessment model, making it difficult to achieve continuous optimization of risk response strategies. Summary of the Invention

[0003] In response to the shortcomings of the existing technology, the present invention provides a real-time early warning and response system for supply chain risks.

[0004] In order to achieve the above object, the technical solution of the present invention is as follows: Real-time supply chain risk warning and response system, including: Data collection module, used to obtain real-time equipment status data of primary suppliers, historical supply interruption records, and real-time logistics route data of backup suppliers; A dynamic weight allocation module generates an alternative priority evaluation model based on preset basic weight parameters and emergency environment parameters, wherein the basic weight parameters include at least logistics timeliness weight, quality consistency weight, and capacity matching weight; An alternative link pre-calculation module generates multi-level alternative paths according to the alternative priority evaluation model and calculates the alternative score of each path; Risk quantification module, used to calculate the risk value of the main supplier based on the main supplier's equipment status data and historical supply interruption records; The instruction generation module, when the risk value of the main supplier exceeds the preset threshold, calls the alternative path with the highest score in the alternative link pre-calculation module and sends a work order change instruction to the production execution system; The dynamic feedback module collects the execution data of the alternative solutions in real time, establishes the deviation relationship between the execution effect index and the pre-calculated parameters, and dynamically modifies the weight allocation rules of the alternative priority evaluation model based on the deviation relationship.

[0005] Compared with the prior art, the present invention has the following beneficial effects: By establishing a pattern matching library between equipment current fluctuation feature vectors and historical shutdown events, traditional qualitative alarms are converted into quantitative risk values ​​of 0-100%, effectively reducing the false alarm rate. The four-dimensional environmental parameter coupling mechanism of the dynamic weight allocation module can exponentially increase the logistics timeliness weight when an orange weather warning is triggered, effectively improving the response speed of alternative solutions; The data acquisition module is used to obtain the main supplier's equipment status and backup supplier's logistics data in real time, and a dynamic weight allocation is combined to generate an alternative priority assessment model. When the risk exceeds the threshold, the optimal alternative plan is automatically triggered and real-time feedback optimization is provided. It has the advantages of real-time capture of dynamic risk factors, intelligent generation of emergency plans and closed-loop optimization decision-making capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] The disclosure of the present invention is described with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. In the drawings, the same reference numerals are used to refer to the same components. Among them: Figure 1 It is a module composition diagram of the present invention; Figure 2 This is a workflow diagram of the dynamic weight module of the present invention; Figure 3 generating a flow chart for the multi-level alternative paths of the present invention; Figure 4 This is a workflow diagram of the risk quantification module of the present invention; Figure 5 This is a workflow diagram of the dynamic feedback module of the present invention. DETAILED DESCRIPTION

[0007] It is easy to understand that according to the technical solution of the present invention, without changing the essential spirit of the present invention, a person skilled in the art can propose a variety of interchangeable structural modes and implementation modes. Therefore, the following specific embodiments and drawings are only exemplary descriptions of the technical solution of the present invention and should not be regarded as the entire invention or as a limitation or restriction of the technical solution of the present invention.

[0008] Application Overview: In traditional supply chain risk management systems, discrete data collection mechanisms can only periodically obtain basic supplier qualification information and are unable to achieve millisecond-level synchronous transmission of equipment operating status parameters, resulting in minute-level delays in identifying abnormal equipment operating conditions. Static weighted assessment models use a fixed ratio of logistics timeliness and quality consistency parameters, without integrating real-time meteorological data and equipment load rate changes. This causes the priority calculation results of alternative suppliers to deviate from actual transportation conditions. The open-loop decision-making architecture lacks a channel for returning execution data. The actual transportation deviation rate and quality fluctuation coefficient of the emergency plan cannot be used to reversely correct the evaluation model parameters, resulting in a continuous decline in model prediction accuracy after the execution of multiple batches of alternative plans.

[0009] For example, in an automotive parts supply scenario, a voltage drop caused abnormal servo motor torque in the main supplier's stamping equipment. Traditional systems could only trigger an alert after the equipment completely shut down. Although the backup supplier met basic production capacity requirements, sudden heavy rain in the area where its logistics route passed led to highway closures. The static evaluation model still recommended this supplier based on the original time-sensitive weighting, and the actual transportation time exceeded the predicted value by three times. When the replacement material arrived, it was found that the dimensional tolerance of the first piece exceeded the standard. The system had not established a correlation between the quality fluctuation data and the evaluation model parameters, and subsequent similar orders continued to recommend this supplier, causing secondary production delays.

[0010] If these issues are not addressed, equipment anomalies will not be detected in their infancy, shortening the window for risk management. Logistics route failures caused by unexpected environmental factors will directly lead to the failure of emergency plan execution. Evaluation models lacking dynamic feedback will continuously output alternative routes that deviate from reality, ultimately leading to the risk of production disruptions. The transmission effect of abnormal operating conditions will spread to secondary suppliers through secondary problems at alternative suppliers, forming a cross-tier supply chain risk network, turning local equipment failures into systemic supply crises.

[0011] When faced with the above problems, this application first addresses the issue of equipment anomaly identification delay, and considers establishing a real-time data acquisition channel to capture millisecond-level equipment operating parameters, while integrating multi-source dynamic environmental data to achieve adaptive adjustment of the risk assessment model. In order to solve the deviation in logistics path failure prediction, it is proposed to build an alternative link pre-calculation mechanism, and screen the optimal emergency plan through a multi-dimensional scoring model. In order to solve the problem of model accuracy degradation caused by the lack of a feedback loop, a reverse correction mechanism for execution effect data is designed to form a dynamic optimization loop. By systematically integrating real-time monitoring, dynamic evaluation, path pre-selection, and closed-loop feedback, a complete link from risk identification to continuous optimization of response strategies is achieved.

[0012] In this regard, Figure 1 As shown, this application proposes: a real-time early warning and response system for supply chain risks, including: A data acquisition module acquires, in real time, equipment state data of the main supplier, historical supply interruption records, and real-time logistics route data of the backup supplier; the data acquisition module refers to a component that acquires, in real time, relevant data of the main supplier and the backup supplier, and can be implemented by using, for example, an Internet of Things sensor to connect an equipment operation monitoring system, a logistics GPS positioning interface call, or an enterprise resource planning system database query, to ensure the timeliness of supply chain state monitoring through real-time synchronization of multi-source data.

[0013] A dynamic weight allocation module generates a substitute priority evaluation model based on preset basic weight parameters and burst environment parameters, and the basic weight parameters at least include a logistics timeliness weight, a quality consistency weight, and a capacity matching degree weight; the dynamic weight allocation module refers to a component that generates an evaluation model according to preset parameters and a burst environment, and can be implemented by using, for example, a multivariate regression algorithm combined with real-time environmental parameter input, to adapt to a burst risk scenario by dynamically adjusting the weight proportion of each evaluation dimension.

[0014] A substitute link precalculation module generates multi-level substitute paths according to the substitute priority evaluation model, and calculates substitute scores of the paths; the substitute link precalculation module refers to a component that generates multi-level substitute paths and calculates scores, and can be implemented by using, for example, a graph theory algorithm to traverse a supplier node topology relationship, to shorten the emergency response time by pre-generating an optimal substitute solution.

[0015] A risk quantification module calculates a main supplier risk value according to equipment state data and historical supply interruption records of the main supplier; the risk quantification module refers to a component that calculates a real-time risk value of the main supplier, and can be implemented by using, for example, a device state data feature extraction and historical interruption record pattern matching algorithm, to provide a trigger threshold basis for decision-making by quantifying a risk probability.

[0016] An instruction generation module calls a substitute path with the highest score in the substitute link precalculation module when the main supplier risk value exceeds a preset threshold, and sends a work order change instruction to a production execution system; the instruction generation module refers to a component that automatically triggers a work order change, and can be implemented by using, for example, an application programming interface to connect the production execution system, to realize an emergency response without manual intervention through automatic transmission of instructions between systems.

[0017] A dynamic feedback module acquires, in real time, substitute solution execution data, establishes a deviation relationship between an execution effect index and precalculation parameters, and dynamically corrects weight allocation rules of the substitute priority evaluation model based on the deviation relationship; the dynamic feedback module refers to a component that corrects weights of the evaluation model, and can be implemented by using, for example, an online training of a machine learning model and an iteration update of weight parameters, to improve the accuracy of a substitute solution through closed-loop feedback of execution effect data.

[0018] The core innovation of this application lies in building a complete closed-loop system that includes data collection, dynamic evaluation, pre-calculation, automatic decision-making and feedback optimization. It solves the response lag problem of traditional solutions in sudden risk scenarios through real-time data fusion and dynamic weighting mechanism. At the same time, it uses execution effect feedback to continuously optimize the evaluation model and form an adaptive supply chain risk response system.

[0019] The working process and principle of this application are that the real-time early warning and response system for supply chain risks realizes risk early warning and response through the collaborative work of multiple modules. The data acquisition module obtains the main supplier's equipment status, historical interruption records and backup supplier logistics route data in real time. The dynamic weight allocation module generates an alternative priority evaluation model based on basic weight parameters and emergency environment parameters. The alternative link pre-calculation module generates multi-level alternative paths according to the evaluation model and calculates the score. The risk quantification module calculates the risk value of the main supplier. When the risk value exceeds the threshold, the instruction generation module calls the highest-scoring alternative path and sends a work order change instruction. The dynamic feedback module collects execution data, establishes deviation relationships and corrects the evaluation model weight rules. Each module realizes a complete link from risk identification to response strategy optimization through data interaction and calculation, thereby improving the real-time and accuracy of supply chain risk management.

[0020] As a preferred embodiment, the solution of this application is specifically implemented as follows: The data acquisition module uses an API interface with the primary supplier's ERP system to obtain real-time equipment operating parameters, including voltage, current, and temperature, with a sampling frequency of 100ms. It also extracts supply disruption records from the past 12 months from a historical database. Logistics route data for backup suppliers is uploaded every five minutes via onboard GPS devices.

[0021] The dynamic weight allocation module presets the basic weights as follows: logistics timeliness 40%, quality consistency 30%, and capacity matching 30%. Emergency environmental parameters include weather warning levels and traffic congestion index.

[0022] The alternative link pre-calculation module generates primary and secondary alternative paths. The primary path is a single backup provider, and the secondary path is a combination of multiple providers.

[0023] The risk quantification module uses a machine learning algorithm to calculate risk values ​​based on equipment parameter fluctuation characteristics and historical interruption patterns.

[0024] The instruction generation module sets the risk threshold at 80%, and automatically triggers alternative plans when it exceeds the threshold.

[0025] The dynamic feedback module records the actual execution time, quality qualification rate and other indicators of the alternative solutions, compares them with the pre-calculated results to generate the deviation rate, and automatically adjusts the corresponding weight when the deviation rate exceeds 10% for three consecutive times.

[0026] Through the above scheme, this application realizes real-time monitoring and rapid response to supply chain risks. Millisecond-level data collection enables equipment anomalies to be identified at an early stage. Dynamic weight distribution takes into account sudden environmental factors and improves the applicability of alternative solutions. The pre-calculation of multi-level alternative paths enhances emergency response capabilities. The risk quantification model combines historical data and real-time parameters to improve the accuracy of early warning. The feedback loop of execution data enables the evaluation model to be continuously optimized and maintains long-term effectiveness. Overall, the system significantly shortens the time from risk identification to response, reduces the impact of supply disruptions on production, and improves the stability and resilience of the supply chain.

[0027] In some of the above-mentioned solutions in this application, when the dynamic weight allocation module generates an evaluation model based on preset parameters, it does not take into account the dynamic impact of environmental variables such as sudden weather changes and policy adjustments on logistics timeliness and quality stability, resulting in the alternative supplier selection mechanism being unable to respond to sudden environmental changes in real time, and the emergency plan is insufficiently matched with actual needs.

[0028] like Figure 2 As shown, the present application further proposes that the dynamic weight allocation module includes a multi-source data coupling unit, which performs the following processing to generate an alternative priority evaluation model: Real-time acquisition of four-dimensional environmental parameters: The logistics timeliness influencing factor extracts the weather warning level of the target area in the next 48 hours from the meteorological warning interface; The quality stability coefficient is obtained through the quality traceability system to obtain the standard deviation of the qualified rate of the backup supplier's latest 10 batches of products; Capacity matching indicators analyze real-time equipment utilization data from the backup supplier's MES system; The order urgency parameter reads the remaining time for delivery of the current order in the ERP system.

[0029] When the weather warning level is ≥ orange warning, the logistics timeliness weight is increased to a predetermined digit, and the priority recalculation of the alternative transportation route is activated at the same time; for every standard deviation interval that the quality stability coefficient decreases, the quality consistency weight increases by a preset percentage, with an upper limit not exceeding 50% of the total weight; for alternative suppliers with equipment utilization rates exceeding 80%, their capacity matching weight decreases according to the overload ratio; the adjusted weights of each dimension are normalized to a total of 100% to generate an alternative priority evaluation model at the current moment.

[0030] The multi-source data coupling unit establishes a real-time data channel with the quality traceability system through a weather warning interface, synchronously updating the weather warning level and pass rate standard deviation every 15 minutes. Equipment utilization data is obtained from the backup supplier's MES system via the OPC protocol, and a sliding window algorithm is used to calculate the average utilization rate over the last hour. The pre-defined digit is set to 1.5 times the base value of the logistics timeliness weight, and the preset percentage is set to increase the quality consistency weight by 5% each time. Normalization calculations use a linear scaling algorithm to ensure that the sum of the weights of each dimension is constant at 100%.

[0031] Specifically, the weather warning interface calls the National Meteorological Administration's API every hour to obtain warning data for the target area for the next 48 hours. When an orange warning or higher is resolved, a mechanism for increasing the weight of logistics timeliness is triggered. For example, the basic logistics timeliness weight is 30%, which increases to 45% after the orange warning is triggered. The quality traceability system stores the inspection data of each batch of products on the blockchain, and the standard deviation is calculated using a moving average method to eliminate data jitter. When equipment utilization is overloaded, the capacity matching weight is dynamically adjusted downward according to the formula [original weight × (1-overload ratio / 20)]. When the equipment utilization rate reaches 90%, the weight drops to 50% of the original value. The remaining time of an order is linked to the production schedule through the ERP system work order number. If the remaining time is less than 24 hours, emergency transportation mode is activated. The adjusted weights are normalized to generate a dynamic evaluation model. The priority of alternative suppliers is updated minute by minute based on real-time environmental parameters to ensure that emergency plans are accurately matched to sudden risk scenarios.

[0032] As a preferred embodiment, the solution of this application is specifically implemented as follows: The dynamic weight assignment module includes a multi-source data coupling unit, which performs the following processing to generate an alternative priority evaluation model: Real-time acquisition of four-dimensional environmental parameters: The logistics timeliness factor extracts the target area's 48-hour weather warning level from the meteorological warning interface. The quality stability coefficient obtains the standard deviation of the backup supplier's product qualification rate for the last 10 batches through the quality traceability system. The capacity matching indicator analyzes real-time equipment utilization data from the backup supplier's MES system. The order urgency parameter reads the remaining delivery time for the current order in the ERP system.

[0033] When the weather warning level reaches or exceeds the orange warning, the logistics timeliness weight is increased to a predetermined digit, and the priority recalculation of the backup transportation route is activated.

[0034] For every standard deviation interval that the quality stability coefficient decreases, the quality consistency weight increases by a preset percentage, with an upper limit not exceeding 50% of the total weight.

[0035] For backup suppliers whose equipment utilization rate exceeds 80%, the capacity matching weight decreases according to the overload ratio.

[0036] The adjusted weights of each dimension are normalized to a total of 100% to generate the alternative priority evaluation model at the current moment, where: normalized weight = (one-way adjusted weight / sum of all phase adjusted weights) × 100%.

[0037] Through the above technical solutions, this application realizes dynamic perception and rapid response to supply chain risks. By acquiring multi-dimensional environmental parameters in real time, the system can promptly capture factors that affect the stability of the supply chain, such as weather changes, quality fluctuations, and production capacity changes. Based on these dynamic data, the system automatically adjusts the weights of each evaluation dimension so that the alternative priority evaluation model can be more in line with the actual situation. This dynamic weight allocation mechanism significantly improves the flexibility and accuracy of risk response, enabling enterprises to make more reasonable decisions in a complex and changing supply chain environment, thereby effectively reducing the risk of supply disruptions and improving the overall stability and reliability of the supply chain.

[0038] In some of the above-mentioned solutions in this application, there is a problem of path simplification in the process of dynamically generating alternative paths, which cannot cover the complex scenario of separation of raw material supply and processing, and the traditional scoring model does not take into account the cost of capacity adjustment, resulting in insufficient economic efficiency of the alternative solutions.

[0039] like Figure 3 As shown, the present application further proposes that the generation process of a multi-level alternative path includes: The first-level alternative path generation screens alternative suppliers that supply the same material type as the primary supplier and whose production capacity gap is smaller than the preset value; The secondary alternative path generation combination includes at least the first supplier and the backup supplier of the second supplier to form a joint alternative plan, in which the first supplier provides the raw materials and the second supplier completes the processing process; The calculation formula for the alternative score of each path is: The replacement score is equal to the logistics timeliness score multiplied by the logistics timeliness weight plus the quality consistency score multiplied by the quality consistency weight plus the capacity matching score multiplied by the capacity matching weight minus the capacity adjustment cost.

[0040] The primary replacement path generation sets the capacity gap threshold as a screening condition, and the preset value is set according to 20% of the historical average daily capacity of the main supplier; the secondary replacement path generation adopts a supplier combination mode, the first supplier selection standard is that the raw material inventory is greater than 150% of the order demand, and the second supplier needs to have the same process standard processing equipment as the main supplier; in the replacement score calculation, the logistics time efficiency score is calculated by the difference ratio of the route passing time and the benchmark time, and the benchmark time is the average transportation time of the logistics line in the past three months; the quality consistency score is inversely converted through the failure rate, and the latest 10 batches of data are obtained from the real-time interface of the supplier quality management system; the capacity matching score is normalized based on the ratio of idle capacity to demand gap, and a full score is obtained when the idle capacity exceeds the demand; the capacity adjustment cost includes the equipment switching cost and the mold replacement cost, the equipment switching cost is calculated by the equipment downtime multiplied by the unit time output value, and the mold replacement cost is calculated by the mold disassembly and installation labor cost.

[0041] Specifically, when generating a primary replacement path, the system screens candidate suppliers whose material types completely match and whose idle capacity meets the main supplier's capacity gap of more than 80%, ensuring that the basic replacement capacity meets the standard. When a single supplier cannot meet the demand, the secondary path generation mechanism is triggered to combine the first supplier with an advantage in raw material inventory and the second supplier with surplus processing capacity to form a joint supply chain link. When calculating the replacement score, the system reflects the transportation stability through the logistics time efficiency score, ensures the material pass rate through the quality consistency score, evaluates the supply capacity adaptation degree through the capacity matching score, and finally deducts the economic cost generated by equipment switching and mold replacement. For example, when the route passing time of a standby supplier is 1.2 times the benchmark time, the logistics time efficiency score is 80 points; if the latest 10 batches of failure rate is 2%, the quality consistency score is 80 points; when the ratio of idle capacity to demand gap is 0.9, the capacity matching score is 90 points; assuming that the capacity adjustment cost is 5000 yuan, the total score comprehensively reflects the technical feasibility and economic rationality. This scheme covers different supply chain scenarios through multiple levels of path, and the quantitative evaluation model realizes the balanced decision of technical indicators and cost factors.

[0042] As a preferred embodiment, the scheme of the application is implemented as follows: The generation process of the multi-level replacement path includes primary replacement path generation and secondary replacement path generation. The primary replacement path generation is realized by screening standby suppliers whose material types are the same as those of the main supplier and whose capacity gap is less than a preset value. For example, when the main supplier A cannot supply a certain electronic component, the system will screen standby suppliers B and C that also produce the electronic component and whose capacity gap is less than 10% as the primary replacement path.

[0043] Secondary replacement paths are generated by combining a set of backup suppliers, including at least a primary supplier and a secondary supplier, to form a joint replacement solution. The primary supplier provides the raw materials, while the secondary supplier completes the processing. For example, if a fully matched primary replacement supplier cannot be found, the system will combine raw material supplier D and processing plant E to form a secondary replacement path, with D providing the raw materials and E completing the processing, thus jointly replacing primary supplier A.

[0044] The alternative score for each path is calculated using the following formula: Replacement score = (Logistics timeliness score × Logistics timeliness weight + Quality consistency score × Quality consistency weight + Capacity matching score × Capacity matching weight) - Capacity adjustment cost; Among them, logistics timeliness score = (1-route travel time / base time) × 100; Quality consistency score = 100 - defective rate of the last 10 batches × 100; Capacity matching score = min (idle capacity / demand gap, 1.0) × 100; The capacity adjustment cost is the weighted sum of the equipment switching cost and the mold replacement cost. For example, for alternative supplier B, assuming its logistics efficiency score is 85, quality consistency score is 95, and capacity matching score is 90, with weights of 0.4, 0.3, and 0.3 respectively, and the capacity adjustment cost is 10, then its replacement score is: (85×0.4+95×0.3+ 90×0.3)-10 =79.

[0045] Through the above technical solution, this application realizes the dynamic generation and scoring of multi-level alternative paths. By combining primary and secondary alternative paths, the flexibility and feasibility of alternative solutions are increased. The alternative scoring formula comprehensively considers multiple dimensions such as logistics timeliness, quality consistency, and capacity matching, and introduces capacity adjustment costs as a balancing factor, making the selection of alternative solutions more comprehensive and reasonable. This method can quickly identify the optimal alternative path and improve the emergency response capability and resilience of the supply chain.

[0046] In some of the above-mentioned solutions in this application, the risk quantification module needs to calculate the risk value based on the equipment status data of the main supplier and the historical supply interruption records, but the following problems exist in the actual operation process: the identification of abnormal equipment signals lacks quantitative standards and cannot be accurately associated with historical failure modes, resulting in insufficient risk prediction accuracy; at the same time, the dynamic correlation between production schedule change delays and order urgency is not considered, which may cause the risk assessment results to deviate from actual business needs.

[0047] like Figure 4As shown, the present application further proposes that the risk quantification module performs the following processing to generate the main supplier risk value: analyze the current fluctuation characteristics in the real-time acquired equipment status data, and when the fluctuation amplitude exceeds a first preset threshold and the duration exceeds a second preset threshold, mark it as an equipment abnormality signal; extract the fault mode that matches the current abnormal signal in the historical supply interruption record, and calculate the pattern matching degree; based on the pattern matching degree and the duration of the equipment abnormal signal, output a real-time risk value of 0-100% through the risk probability model.

[0048] The current fluctuation characteristics are analyzed using a sliding window variance calculation method. The preset threshold is set between 20% and 30% of the device's rated current, and the duration threshold is set between 5 and 15 minutes. Pattern matching is calculated using a cosine similarity algorithm, comparing the current waveform feature vectors corresponding to historical fault patterns. A risk probability model is constructed using a Bayesian network. The pattern matching and duration are input into a conditional probability table, which outputs a risk probability value.

[0049] Specifically, the equipment status data acquisition unit collects current data at a frequency of once per second and eliminates instantaneous interference through moving average filtering. When the current fluctuation amplitude exceeds 25% of the rated current for three consecutive sampling cycles and lasts for 10 minutes, the equipment abnormality signal is triggered. At this time, the fault mode library is called for matching. For example, the mode number F023 that caused a shutdown with similar fluctuation characteristics in the historical records has a similarity of 0.85. According to the basic risk probability of 45% corresponding to this mode, combined with the current abnormality lasting 12 minutes, the real-time risk value calculated by the risk probability model is increased to 58%. When the risk value exceeds the preset threshold of 50%, the subsequent alternative path call process is triggered. This process achieves dynamic and accurate assessment of the risk value by quantifying the correlation between the physical characteristics of the abnormal signal and the historical pattern.

[0050] As a preferred embodiment, the solution of this application is specifically implemented as follows: The risk quantification module performs the following processing to generate the primary supplier risk value: First, analyze the current fluctuation characteristics in the real-time device status data. Specifically, by collecting device current data, the standard deviation of the current values ​​is calculated. If the standard deviation exceeds a first preset threshold of 0.5A and persists for more than a second preset threshold of 10 minutes, it is marked as a device abnormality signal.

[0051] Next, the system extracts fault patterns from historical power outage records that match the current abnormal signal. For example, by comparing the current current fluctuation pattern with the historical fault current waveform, the similarity is calculated. If the similarity exceeds 80%, the fault pattern is determined to match, and the pattern matching degree is calculated.

[0052] Finally, based on the pattern matching degree and the duration of the device abnormal signal, a risk probability model outputs a real-time risk value of 0-100%. Specifically, the risk probability model uses a logistic regression algorithm, with pattern matching degree and abnormality duration as input variables, to output a risk probability value of 0-100%.

[0053] Through the above technical solution, this application achieves real-time quantitative assessment of key supplier risks. This enables the system to promptly capture equipment operating anomalies and, combined with historical data, predict risks, improving the real-time and accuracy of risk identification. Furthermore, by outputting specific risk probability values, it provides a quantitative basis for subsequent risk response decisions, helping to improve the accuracy and effectiveness of supply chain risk management.

[0054] In some of the above-mentioned solutions in this application, the dynamic correlation between production schedule change delay and order urgency is not integrated in the calculation process of the main supplier's risk value, resulting in insufficient matching between the risk quantification results and actual business needs, which may cause excessive or delayed emergency response.

[0055] The present application further proposes that the risk quantification module includes a multi-source data fusion unit, which performs the following operations: Synchronously obtain production schedule data from the main supplier and detect the delay in schedule changes when equipment abnormality signals occur; Associate the order urgency tag of the procurement system and calculate the risk value compensation coefficient: The compensation coefficient is equal to the ratio of the scheduling delay duration to the remaining order delivery duration, and the maximum value does not exceed 1; The final risk value is equal to the basic risk value multiplied by one and the compensation coefficient, and the maximum value does not exceed one hundred percent.

[0056] Among them, the synchronous acquisition of the main supplier's production schedule data is achieved through the interface docking with the manufacturing execution system. The starting point for detecting the delay time of the schedule change is the time difference between the time when the equipment abnormality signal is triggered and the time when the schedule adjustment instruction is actually issued. The order urgency label is divided according to the delivery remaining time threshold interval preset in the procurement system, including three types of status: urgent, regular, and loose. The calculation of the compensation coefficient adopts the ratio function constraint. When the scheduling delay time exceeds the remaining time of the order delivery, the compensation coefficient takes the maximum value of one. The adjustment mechanism of the final risk value is realized through linear superposition. The product of the basic risk value and the compensation coefficient partially reflects the amplifying effect of scheduling delay on the pressure of order delivery.

[0057] Specifically, when an equipment anomaly signal is triggered, the system automatically retrieves the current production schedule data of the primary supplier and records the actual response time from the occurrence of the anomaly to the schedule adjustment as the schedule change delay. Simultaneously, it accesses the procurement system to obtain the remaining delivery time data for the current order, establishing a quantitative correlation between the delay time and delivery pressure. The compensation coefficient is calculated by dividing the scheduling delay time by the remaining delivery time of the order. This ratio is constrained to a closed interval between zero and one to ensure the controllability of the risk value increase. The final risk value is generated through a linear combination of the base risk value and the compensation coefficient. When the base risk value is 50% and the compensation coefficient is 1, the final risk value reaches its maximum value of 100%. This mechanism allows high-risk orders to trigger warning thresholds more quickly when scheduling response delays occur, ensuring that alternative plans are initiated in a timely manner while avoiding over-response to low-risk orders.

[0058] As a preferred embodiment, the solution of this application is specifically implemented as follows: The risk quantification module includes a multi-source data fusion unit. The multi-source data fusion unit performs the following operations: First, the system synchronously acquires production schedule data from the primary supplier. When a device anomaly signal is detected, the system immediately measures the delay in scheduling changes. For example, if the original production schedule calls for 100 units in 8 hours, but an equipment anomaly causes actual production to drop to 10 units per hour, the system will calculate a 4-hour scheduling delay.

[0059] Next, we link the order urgency tag in the procurement system to calculate the risk compensation coefficient. The compensation coefficient is calculated as follows: Compensation coefficient = min(schedule delay duration / order delivery remaining time, 1.0). For example, if the order delivery remaining time is 8 hours and the scheduling delay is 4 hours, the compensation coefficient is 0.5.

[0060] Finally, calculate the final risk value. The formula is: Final Risk Value = Base Risk Value × (1 + Compensation Factor). It's important to note that the maximum value of the Final Risk Value cannot exceed 100%. For example, if the Base Risk Value is 60% and the Compensation Factor is 0.5, the Final Risk Value is 90%.

[0061] Through the above technical solution, this application achieves precise quantification of supply chain risks. By integrating production scheduling data and order urgency information, the system can more comprehensively assess the impact of equipment anomalies on actual production and order delivery. This dynamic risk assessment method significantly improves the accuracy and timeliness of risk warnings, enabling companies to make more rapid and effective response decisions, thereby reducing losses caused by supply disruptions.

[0062] In some of the above-mentioned schemes in this application, the calculation of the main supplier risk value is based on the matching degree between the equipment abnormality signal and the historical failure mode, but the correlation between the duration of the equipment abnormality and the risk probability is only processed through a linear relationship, which fails to accurately reflect the cumulative effect of continuous abnormalities on the risk of supply chain disruption, resulting in insufficient response sensitivity and accuracy of the risk prediction model to sudden equipment failures.

[0063] This application further proposes the construction of a risk probability model, including: The historical data analysis submodule calculates the actual shutdown probability corresponding to each failure mode and establishes a failure mode library; The real-time prediction submodule calculates the similarity between the current abnormal signal and the historical fault pattern through a sliding time window. The similarity is obtained by vector dot product and norm ratio. When the similarity exceeds the preset matching threshold, the shutdown probability of the corresponding historical fault pattern is used as the basic risk value. The real-time prediction submodule adds a dynamic attenuation factor to increase the risk value according to an exponential function when the duration of the device abnormal signal reaches a preset critical point.

[0064] The historical data analysis submodule statistically analyzes three years of equipment failure event data to establish a database containing twelve typical failure modes. Each mode is associated with a current fluctuation feature vector and an average downtime probability. The real-time prediction submodule collects equipment current data every five minutes to generate an eight-dimensional fluctuation feature vector. This vector is then compared with the historical pattern vector using cosine similarity calculations. When the similarity exceeds a threshold of 0.85, a pattern match is triggered. A dynamic attenuation factor is activated after an anomaly persists for more than four hours, increasing the risk value exponentially. , where when the duration exceeds the critical point, the risk value growth rate increases by 200%-500%.

[0065] Specifically, when the equipment current fluctuation amplitude exceeds the baseline value by 20% for 30 consecutive minutes, a feature vector containing amplitude, frequency, and duration is generated. This feature vector is then compared with historical vectors in the fault pattern library using a sliding window mechanism. If a match is found for the 85% downtime probability corresponding to the fourth type of failure mode, the initial risk value is set to 85%.

[0066] When the abnormal state lasts for the fifth hour, the risk value calculated by applying the dynamic attenuation factor increases to: Because the system sets a maximum risk value of 100%, the final output risk value is capped at 100%. This mechanism establishes a nonlinear relationship between the duration of equipment anomalies and the risk probability, enabling the system to accurately reflect the cumulative effects of risk when responding to intermittent equipment failures while avoiding premature emergency response triggering due to misjudgment at a single point in time.

[0067] As a preferred embodiment, the solution of this application is specifically implemented as follows: The risk probability model is constructed using a historical data analysis submodule and a real-time prediction submodule. The historical data analysis submodule calculates the actual downtime probability corresponding to each failure mode and establishes a failure mode library. The real-time prediction submodule calculates the similarity between the current abnormal signal and historical failure modes using a sliding time window.

[0068] The similarity calculation formula is: Similarity = Σ(current fluctuation feature vector · historical pattern vector) / (||current vector|| × ||historical vector||). When the similarity exceeds the preset matching threshold, the downtime probability of the corresponding historical failure mode is used as the basic risk value.

[0069] The real-time prediction submodule also includes a dynamic attenuation factor. When the duration of an abnormal device signal reaches a preset threshold, the risk value increases exponentially. For example, you could set the threshold to 2 hours. If the abnormal signal persists for more than 2 hours, the risk value increases by 200%-500%.

[0070] Through the above technical solution, this application achieves a precise quantitative assessment of supplier equipment failure risks. By establishing a historical failure pattern library and calculating the similarity of real-time abnormal signals, potential equipment failure risks can be quickly identified. The introduction of a dynamic attenuation factor further improves the timeliness of risk assessment, avoiding the lag problem of risk assessment in traditional methods. This real-time, dynamic risk quantification method provides a reliable basis for companies to take timely risk response measures, effectively reducing the risk of supply chain disruptions.

[0071] In some of the above-mentioned solutions in this application, the dynamic feedback module dynamically corrects the alternative priority evaluation model through the deviation relationship between the execution effect index and the pre-calculated parameters. However, in the actual implementation process, there is an uncontrollable deviation between the logistics time prediction value and the actual transportation time, and the quality fluctuations of alternative suppliers may not be captured in time, resulting in a lack of accurate data support for the weight adjustment.

[0072] This application further proposes the deviation relationship between the execution effect index and the pre-calculated parameters, including the logistics timeliness deviation rate and the quality fluctuation coefficient; Among them, the logistics time deviation rate = (actual transportation time - predicted time) / predicted time × 100%.

[0073] The logistics time deviation rate is calculated by collecting coordinate data in real time from the transport vehicle's GPS terminal. This data is then spatiotemporally matched against pre-calculated route nodes. Deviation rate calculation is triggered when the actual transport time exceeds the predicted value. The quality fluctuation coefficient is calculated by obtaining incoming material inspection reports from alternative suppliers through the MES system, and the standard deviation of the batch pass rate is used as a quantitative basis. Every 5% increase in the logistics time deviation rate triggers a weight iteration, and every 0.1 standard deviation decrease in the quality fluctuation coefficient triggers a weight correction instruction.

[0074] Specifically, the logistics timeliness deviation rate calculation unit obtains the location coordinates of the transport vehicle every 15 minutes, compares the actual driving path with the pre-calculated route, and activates the deviation rate calculation engine when it detects that the detour distance exceeds the preset threshold. The quality fluctuation coefficient monitoring module scans the quality traceability database in real time and performs a sliding window analysis on the incoming material inspection data of alternative suppliers. When the fluctuation range of the defective rate of three consecutive batches exceeds 2 times the historical standard deviation, an abnormal quality fluctuation coefficient event is automatically generated. The deviation relationship model normalizes the logistics timeliness deviation rate and the quality fluctuation coefficient, generates a two-dimensional vector input weight distribution algorithm, and optimizes the weight parameters through the gradient descent method, so that the error rate between the actual execution effect of the alternative path scoring model and the pre-calculated result is controlled within 3%.

[0075] As a preferred embodiment, the solution of this application is specifically implemented as follows: The deviation relationship between execution performance indicators and pre-calculated parameters includes the logistics time deviation rate and the quality fluctuation coefficient. The logistics time deviation rate is calculated as the difference between actual transportation time and predicted transportation time. Specifically, the logistics time deviation rate = (actual transportation time - predicted transportation time) / predicted transportation time × 100%.

[0076] For example, if the predicted transport time for a batch of goods is 10 hours and the actual transport time is 12 hours, then the logistics time deviation rate = (12-10) / 10×100% = 20%. This means that the actual transport time is 20% longer than expected.

[0077] The quality fluctuation coefficient can be measured by calculating the standard deviation of the quality test results of consecutive batches of products from alternative suppliers. .

[0078] Through the quality fluctuation coefficient, the system can promptly detect abnormal situations during the execution of alternative plans and provide data support for the dynamic optimization of the alternative priority evaluation model.

[0079] Through the above-mentioned technical solution, this application can achieve a quantitative assessment of the execution effect of alternative solutions, providing an objective basis for dynamically adjusting the alternative priority assessment model. As a result, the system can continuously optimize the decision-making model based on actual execution, improving the accuracy and efficiency of responding to supply chain risks. Furthermore, by establishing a correlation between execution effect and pre-calculated parameters, the system can continuously learn and improve, enhancing its adaptability to complex and changing supply chain environments.

[0080] In some of the above-mentioned schemes in this application, the dynamic feedback module dynamically corrects the weight distribution rules of the replacement priority evaluation model through the deviation relationship between the execution effect index and the pre-calculated parameters. However, in actual applications, the dynamic adjustment mechanism of the logistics timeliness deviation rate has a response delay problem, which cannot reflect the actual execution status of the transportation path in a timely manner, resulting in the weight correction lagging behind the real-time logistics status changes.

[0081] like Figure 5 As shown, the present application further proposes that the dynamic feedback module includes a real-time data synchronization unit and a weight iteration unit.

[0082] The real-time data synchronization unit obtains the location coordinates from the transport vehicle GPS terminal every T minutes and matches and verifies them with the pre-calculated path nodes.

[0083] The weight iteration unit automatically reduces the priority of the corresponding transport route's timeliness weight when the logistics timeliness deviation rate exceeds the preset tolerance threshold for N consecutive times. The new timeliness weight is calculated by multiplying the original weight by the proportional adjustment item of the deviation rate and the maximum tolerance threshold and the configurable risk sensitivity coefficient.

[0084] The real-time data synchronization unit collects the geographic location information of transport vehicles at fixed intervals and verifies the consistency between the actual route and the pre-calculated route using a coordinate matching algorithm. The weight iteration unit uses a continuous trigger mechanism to initiate weight adjustment calculations when the cumulative number of times the logistics timeliness deviation rate exceeds the tolerance threshold reaches N. The mathematical expression of the new timeliness weight is a composite function of the original weight, the deviation rate ratio, and the risk sensitivity coefficient, where the risk sensitivity coefficient is negatively correlated with the reliability level of the transport route.

[0085] Specifically, the real-time data synchronization unit establishes a time series dataset of the transport trajectory through periodic GPS data collection, and uses spatial geo-fencing technology to compare the degree of deviation between the actual position and the pre-calculated path nodes. When the system detects that the path deviation exceeds the standard within N consecutive collection cycles, the weight iteration unit activates the time-sensitive weight recalculation process. The reliability level is introduced as a damping coefficient in the calculation process. This coefficient is derived based on the statistics of historical deviation records and effectively suppresses the excessive interference of occasional abnormal data on weight adjustment. The risk sensitivity coefficient is set to the complement of the reliability level to ensure that the weight adjustment range of high-reliability routes is smaller than that of low-reliability routes, realizing a differentiated dynamic optimization mechanism.

[0086] As a preferred embodiment, the solution of the present application is specifically implemented as follows: the real-time data synchronization unit obtains the longitude and latitude coordinates from the GPS terminal of the transport vehicle every 15 minutes, and matches and verifies them with the pre-calculated path nodes, where the path nodes include the coordinates of key hubs in the planned route. When the logistics timeliness deviation rate exceeds the preset 10% tolerance threshold for three consecutive times, the weight iteration unit automatically triggers the timeliness weight adjustment mechanism, and the new timeliness weight is calculated by multiplying the original weight by (1-deviation rate / 10%) and the risk sensitivity coefficient. The risk sensitivity coefficient is dynamically determined based on the transport timeliness credibility matrix, which calculates the route reliability level based on historical deviation rate data, where the reliability level is a normalized value of 0-1. For example, when the historical deviation rate of a route is less than 5%, its reliability level is set to 0.9, and the risk sensitivity coefficient is adjusted to 0.1. Using this calculation method, when a route with an original timeliness weight of 40% experiences a 12% deviation rate, the new weight is adjusted to 40%×[1-(12% / 10%)×(1-0.9)]=38.08%, thus achieving dynamic degradation of timeliness priority.

[0087] Through the above technical solutions, this application effectively solves the problem of rigid time evaluation of transportation routes in traditional supply chain systems. Through real-time data synchronization and deviation rate triggering mechanism, the weight distribution strategy can be dynamically corrected. Specifically in the logistics execution scenario, when the actual transportation time continues to deviate from the predicted value, the system automatically reduces the time evaluation priority of the route to avoid the failure of alternative execution due to the decline in route reliability. At the same time, the reliability level is introduced as a damping coefficient to prevent occasional deviations from excessively affecting the weight adjustment, ensuring the stability and adaptability of the dynamic feedback process.

[0088] In some of the above-mentioned solutions of this application, when adjusting weights based on the real-time logistics time deviation rate, the historical reliability differences of transportation routes are not taken into account, resulting in a single deviation event that may excessively affect the weight calculation, causing insufficient stability of the evaluation model.

[0089] This application further proposes to establish a transport time reliability matrix and calculate the route reliability level based on the deviation rate history: ; The reliability level is used as the damping coefficient for timeliness weight adjustment, and the risk sensitivity coefficient is set to 1-reliability level; the calculation formula for the new timeliness weight is updated to the original weight multiplied by [1-(deviation rate / maximum tolerance threshold) multiplied by (1-reliability level)].

[0090] The transport time reliability matrix is ​​constructed by storing logistics time deviation rate data for M consecutive cycles, where M is a configurable integer, and each cycle corresponds to a preset time period. Route reliability levels are divided into high, medium, and low levels based on weighted statistics of the frequency and magnitude of historical deviation rates, with corresponding values ​​of 0.8, 0.5, and 0.2, respectively. The damping coefficient maps the reliability level value to the adjustment formula, creating a buffer against sudden deviation events. The risk sensitivity coefficient is inversely correlated with the reliability level, ensuring that weight adjustments for low-reliability routes are suppressed. After the updated formula, the impact of the deviation rate on the weight is negatively correlated with the route's historical performance.

[0091] Specifically, when the historical reliability level of transport route A is high, its corresponding risk sensitivity coefficient is automatically adjusted to 0.2. If the current logistics timeliness deviation rate reaches 80% of the maximum tolerance threshold, when calculating the new timeliness weight, the actual adjustment range is [1-(0.8×(1-0.8))] of the original weight, that is, 84% of the original weight is retained. On the contrary, for low-reliability route B, under the same deviation rate, the adjustment range reaches [1-(0.8×(1-0.2))] of the original weight, that is, 36% of the original weight is retained. This mechanism effectively distinguishes occasional deviations from systematically unreliable routes, prevents high-reliability routes from being excessively downgraded due to single anomalies, and accelerates the elimination of persistently inefficient paths. By introducing the dimension of historical data, the weight iteration process has the ability to perform time series analysis, thereby improving the anti-interference ability and long-term decision consistency of the alternative evaluation model.

[0092] As a preferred embodiment, the solution of the present application is specifically implemented as follows: During the logistics transportation process, the GPS terminal of the transport vehicle synchronizes the real-time coordinate data to the dynamic feedback module every 15 minutes. The actual driving trajectory of the vehicle is compared with the pre-calculated path node through the path matching algorithm. When it is detected that the logistics timeliness deviation rate of a certain section exceeds 20% for three consecutive times, the system automatically triggers the weight iterative calculation. Based on the deviation records of the transport route in the past 30 days, a credibility matrix containing time dimension and space dimension is constructed, where the reliability level is divided into five levels of AE according to the historical deviation frequency and amplitude. The transport route identified as Class C reliability is taken as the adjustment object, and its risk sensitivity coefficient is set to 0.4. In the weight update stage, the original logistics timeliness weight is 35%, and the new timeliness weight calculated according to the correction formula is 35%×[1-(25% / 30%)×(1-0.6)], where the maximum tolerance threshold is set to 30%, and the damping coefficient corresponding to the reliability level is 0.6.

[0093] Through the above technical solution, this application effectively solves the problem of inaccurate weight adjustment caused by the lack of transportation route reliability assessment in traditional supply chain management systems. By establishing a multi-dimensional credibility matrix and introducing reliability levels as damping coefficients, the dynamic adjustment of timeliness weights can not only reflect the urgency of current transportation deviations, but also take into account the stability characteristics of historical transportation data. This avoids excessive weight adjustments caused by single abnormal events, ensures the rationality of decision-making of alternative solution evaluation models in different transportation scenarios, and improves the matching accuracy of emergency plans with actual logistics conditions.

[0094] In some of the aforementioned solutions, the dynamic feedback module adjusts the weight distribution rules based on the logistics timeliness deviation rate. However, this does not address the risk of quality fluctuations from alternative suppliers transmitting to the upstream supply chain. When an alternative supplier exhibits quality defects, existing solutions are unable to effectively identify the root cause of the defect and update the alternative path, causing quality risks to spread upward along the supply chain.

[0095] This application further proposes a quality feedback loop unit that performs the following processing: Scan incoming material inspection reports in the MES system to extract the first-article qualification rate of alternative suppliers; When the first-article qualified rate is lower than the preset threshold of the supplier's historical level, a temporary downward adjustment instruction of the quality consistency weight is triggered; If the qualified rate of P consecutive batches returns to the benchmark value, the temporary reduction will be automatically lifted and the weight coefficient will be compensated; Establish a quality risk transmission model. When quality fluctuations of alternative suppliers are detected: trace the upstream raw material sources associated with the supplier; Analyze the correlation between quality defect patterns and raw material batches; when the correlation exceeds the preset threshold, automatically transmit the risk to the second-tier supplier node and update the alternative path score; Wherein, P is a configurable positive integer.

[0096] The quality feedback loop unit includes three key processing links: a first-article pass rate monitoring mechanism, a weighted dynamic compensation mechanism, and a quality risk transmission model. The first-article pass rate monitoring mechanism obtains the first inspection data of alternative suppliers in real time through the MES system interface, and sets 80% of the historical pass rate average as the temporary downward adjustment threshold. The weighted dynamic compensation mechanism uses time series judgment logic. When the pass rate of three consecutive production batches (P=3) returns to the baseline value, a weighted recovery operation is performed, and the recovery amount is 120% of the original downward adjustment range. The quality risk transmission model has a built-in defect pattern matching algorithm, which calculates the cosine similarity between the current quality defect characteristics and the abnormal records in the raw material batch database, and the correlation threshold is set to 0.75.

[0097] Specifically, the workflow of the quality feedback loop unit is divided into two stages: quality fluctuation detection and risk transmission control. In the detection stage, the system automatically captures the first-article inspection report in the MES every two hours and calculates the deviation between the current qualified rate and the supplier's average value over the past 30 days. When the deviation exceeds 20%, the weight reduction instruction is triggered, and the quality consistency weight is immediately reduced by 15 percentage points. In the risk transmission stage, the system reversely tracks batches of materials with quality defects and links them to secondary raw material suppliers through the supplier's BOM table. If the defect pattern analysis shows that the temperature fluctuation of the raw material heat treatment is strongly correlated with the current defective rate (correlation coefficient ≥ 0.8), the alternative path score of the secondary supplier will be automatically reduced by 20%, and a supplier switching recommendation will be generated. This dual control mechanism realizes closed-loop management from terminal quality anomalies to upstream risk nodes, ensuring that the adjustment of alternative solutions covers both direct suppliers and indirect supply chain nodes.

[0098] As a preferred embodiment, the solution of the present application is specifically implemented as follows: the quality feedback loop unit is configured to perform quality monitoring and weight adjustment operations. The incoming material inspection report of the MES system is scanned in real time, and the first-article qualification rate of the alternative supplier is automatically extracted through the database interface. When it is detected that the first-article qualification rate of a supplier drops by more than 15 percentage points compared with its historical three-month average level, a temporary reduction instruction for the quality consistency weight is triggered, and the quality weight coefficient of the supplier in the alternative priority evaluation model is reduced by 30%. If the incoming material qualification rate of the supplier for five consecutive batches is restored to the historical benchmark value within ±2%, the temporary reduction instruction is automatically lifted and the weight coefficient is compensated by 10%. The quality risk transmission model is further constructed to include an upstream raw material traceability mechanism. When the dimensional defective rate of the injection molding supplier increases abnormally, the batch number of the resin raw material it purchases is correlated and analyzed, and the raw material melt index is found to exceed the standard through spectral detection. At this time, the quality fluctuation coefficient of the raw material supplier is updated to the secondary alternative path scoring model, triggering the recalculation of the secondary supplier node weight.

[0099] Through the above technical solutions, this application realizes dynamic closed-loop management of supply chain quality risks. The quality consistency weight can be adjusted in real time according to the actual incoming material situation, avoiding emergency decision-making deviations caused by a fixed scoring mechanism. The automatic traceability mechanism of the quality defect transmission path effectively identifies the risks associated with multi-level suppliers and prevents local quality problems from causing cascading failures in the supply chain. The continuous interactive optimization of execution feedback data and model parameters enhances the adaptability of alternative solutions to actual production conditions, significantly improving the accuracy and timeliness of supply chain disruption response strategies.

[0100] The technical scope of the present invention is not limited to the contents of the above description. Those skilled in the art can make various deformations and modifications to the above embodiments without departing from the technical idea of ​​the present invention, and these deformations and modifications should all fall within the protection scope of the present invention.

Claims

1. Real-time supply chain risk warning and response system, featuring: include: Data collection module, used to obtain real-time equipment status data of primary suppliers, historical supply interruption records, and real-time logistics route data of backup suppliers; A dynamic weight allocation module generates an alternative priority evaluation model based on preset basic weight parameters and emergency environment parameters, wherein the basic weight parameters include at least logistics timeliness weight, quality consistency weight, and capacity matching weight; An alternative link pre-calculation module generates multi-level alternative paths according to the alternative priority evaluation model and calculates the alternative score of each path; Risk quantification module, used to calculate the risk value of the main supplier based on the main supplier's equipment status data and historical supply interruption records; The instruction generation module, when the risk value of the main supplier exceeds the preset threshold, calls the alternative path with the highest score in the alternative link pre-calculation module and sends a work order change instruction to the production execution system; The dynamic feedback module collects the execution data of the alternative solutions in real time, establishes the deviation relationship between the execution effect index and the pre-calculated parameters, and dynamically modifies the weight allocation rules of the alternative priority evaluation model based on the deviation relationship.

2. The real-time supply chain risk warning and response system according to claim 1 is characterized by: The dynamic weight allocation module includes: a multi-source data coupling unit, which performs the following processing to generate an alternative priority evaluation model: Real-time acquisition of four-dimensional environmental parameters: logistics timeliness influencing factor, quality stability coefficient, production capacity matching index, and order urgency parameter; When the weather warning level is ≥ orange warning, the logistics timeliness weight is increased to a predetermined digit, and the priority recalculation of the backup transportation route is activated; For every standard deviation interval that the quality stability coefficient decreases, the quality consistency weight increases by a preset percentage, with an upper limit not exceeding 50% of the total weight; For backup suppliers whose equipment utilization rate exceeds 80%, their capacity matching weight decreases according to the overload ratio; The adjusted weights of each dimension are normalized to a total of 100% to generate the current substitution priority evaluation model, where: Normalized weight = (one-way adjusted weight / sum of all phase adjusted weights) × 100%.

3. The real-time supply chain risk warning and response system according to claim 2, characterized in that: The generation process of the multi-level alternative path includes: First-level alternative path generation: Screening of backup suppliers with the same material type as the primary supplier and whose production capacity gap is smaller than the preset value; Secondary alternative path generation: a combination of alternative suppliers including at least the first supplier and the second supplier is used to form a joint alternative solution, where the first supplier provides the raw materials and the second supplier completes the processing steps; The calculation formula for the alternative score of each path is: Replacement score = (Logistics timeliness score × Logistics timeliness weight + Quality consistency score × Quality consistency weight + Capacity matching score × Capacity matching weight) - Capacity adjustment cost; Among them, logistics timeliness score = (1-route travel time / base time) × 100; Quality consistency score = 100 - defective rate of the last 10 batches × 100; Capacity matching score = min (idle capacity / demand gap, 1.0) × 100; The capacity adjustment cost is the weighted sum of the equipment switching cost and the mold replacement cost.

4. The real-time supply chain risk warning and response system according to claim 1 is characterized by: The risk quantification module performs the following processing to generate the main supplier risk value: Analyze the current fluctuation characteristics in the real-time acquired device status data, and mark it as a device abnormality signal when the fluctuation amplitude exceeds a first preset threshold and the duration exceeds a second preset threshold; Extract the fault pattern that matches the current abnormal signal from the historical supply interruption records and calculate the pattern matching degree; Based on the pattern matching degree and the duration of the device abnormal signal, a real-time risk value of 0-100% is output through the risk probability model.

5. The real-time supply chain risk warning and response system according to claim 4 is characterized by: The risk quantification module includes: The multi-source data fusion unit performs the following operations: Synchronously obtain production schedule data from the main supplier and detect the delay in schedule changes when equipment abnormality signals occur; Associate the order urgency tag of the procurement system and calculate the risk value compensation coefficient: Compensation coefficient = min(scheduling delay duration / order delivery remaining duration, 1.0); The final risk value = basic risk value × (1 + compensation coefficient), and the maximum value shall not exceed 100%.

6. The real-time supply chain risk warning and response system according to claim 4 is characterized by: The construction of the risk probability model includes: The historical data analysis submodule calculates the actual shutdown probability corresponding to each failure mode and establishes a failure mode library; The real-time prediction submodule calculates the similarity between the current abnormal signal and the historical fault mode through a sliding time window: Similarity = Σ(current fluctuation feature vector·historical pattern vector) / (||current vector||×||historical vector||); When the similarity exceeds the preset matching threshold, the downtime probability of the corresponding historical failure mode is called as the basic risk value; The real-time prediction submodule also includes: Dynamic attenuation factor: When the duration of abnormal device signals reaches a preset critical point, the risk value increases exponentially: .

7. The real-time supply chain risk warning and response system according to claim 1 is characterized by: The deviation relationship between the execution effect index and the pre-calculated parameters includes the logistics timeliness deviation rate and the quality fluctuation coefficient; Among them, logistics time deviation rate = (actual transportation time - predicted time) / predicted time × 100%; 。 8. The real-time supply chain risk warning and response system according to claim 7, characterized in that: The dynamic feedback module includes: Real-time data synchronization unit, which obtains location coordinates from the transport vehicle GPS terminal every T minutes and matches and verifies them with the pre-calculated path nodes; The weight iteration unit automatically reduces the priority of the corresponding transport route’s timeliness weight when the logistics timeliness deviation rate exceeds the preset tolerance threshold for N consecutive times. The calculation formula is: New timeliness weight = original weight × (1-deviation rate / maximum tolerance threshold) × configurable risk sensitivity coefficient; Wherein, T and N are both configurable positive integers.

9. The real-time supply chain risk warning and response system according to claim 8, characterized in that: The processing process of the weight iteration unit includes: Establish a transport time reliability matrix and calculate the route reliability level based on the historical deviation rate records: ； The reliability level is used as the damping coefficient for timeliness weight adjustment, and the configurable risk sensitivity coefficient is set to: risk sensitivity coefficient = 1-reliability level; The calculation formula of the new timeliness weight is updated as follows: New timeliness weight = original weight × [1-(deviation rate / maximum tolerance threshold) × (1-reliability level)].

10. The supply chain risk real-time warning and response system according to claim 1, characterized in that: The dynamic feedback module also includes: The quality feedback loop unit performs the following processing: Scan incoming material inspection reports in the MES system to extract the first-article qualification rate of alternative suppliers; When the first-article qualified rate is lower than the preset threshold of the supplier's historical level, a temporary downward adjustment instruction of the quality consistency weight is triggered; If the qualified rate of P consecutive batches returns to the benchmark value, the temporary reduction will be automatically lifted and the weight coefficient will be compensated; Establish a quality risk transmission model. When quality fluctuations of alternative suppliers are detected: Trace the upstream raw material sources associated with the supplier; Analyze the correlation between quality defect patterns and raw material batches; When the correlation exceeds the preset critical value, the risk is automatically transmitted to the second-tier supplier node and the alternative path score is updated; Wherein, P is a configurable positive integer.

Images