Internet of Things platform supply chain management system based on multi-source data analysis

Abstract

The invention relates to the technical field of Internet of Things platforms, in particular to an Internet of Things platform supply chain management system based on multi-source data analysis, which comprises an acquisition module, a consumption calculation module, a period calculation module, a state determination module, a generation module, a deviation calculation module and a correction module. According to the method, multi-dimensional operation data such as the vehicle operation mileage, the operation duration and the departure shift are subjected to weighted normalization processing to obtain the operation intensity, the unit cycle consumption is dynamically corrected according to the operation intensity, and meanwhile, the inventory supportable cycle is calculated according to the ratio of the total inventory to the predicted cycle consumption; and comparing the average delivery duration with the average delivery duration of the suppliers, so that intelligent collaborative allocation of the key spare parts of the new energy bus station can be realized. According to the method, the problems of demand response lag, low running intensity change matching degree and low matching speed in a dynamic consumption scene due to excessive dependence on historical periodic prediction are effectively solved.

Description

A Supply Chain Management System Based on Multi-Source Data Analysis (IoT Platform) Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) platform technology, and in particular to an IoT platform supply chain management system based on multi-source data analysis. Background Technology

[0002] As urban public transportation systems continue to evolve towards digitalization, networking, and refined operation, the scale of operations is expanding and the workload is becoming increasingly complex. Shifting resource allocation from experience-driven to data-driven approaches is becoming an inevitable trend. However, with the coexistence of multi-node collaborative operation and high-frequency dynamic demand, factors such as demand fluctuations, delivery cycle differences, and unbalanced inventory structures combine to create dual pressures on cost control and operational stability. Achieving dynamic matching of supply and demand in an uncertain environment has become a core challenge in the construction of intelligent operation systems.

[0003] Chinese Patent Publication No. CN118278861A discloses a multi-layered supply chain network inventory management method. This method first represents the logistics distribution relationship of the supply chain network as a connected directed graph, where goods can only flow unidirectionally from the upstream layer to the downstream layer. The method predicts the future demand sequence for each period based on the historical consumption data of the downstream layer nodes, forming an external demand forecast column vector for all nodes. Using a matrix representing the replenishment lead time and order allocation weights between upstream and downstream nodes, this external demand forecast column vector is converted into an internal demand forecast sequence for all nodes in the entire network for each future period. Each node determines the total amount of goods it needs to order from upstream suppliers in the current inventory period based on its own demand forecast sequence and sends specific order quantities to its respective upstream supplier nodes according to preset order allocation weights.

[0004] Therefore, the existing technology has the following problems: it relies on periodic demand forecasts based on historical consumption data as the main basis for decision-making, without fully considering the immediate impact of changes in operating load on the consumption rhythm, which can easily lead to forecast lag when demand fluctuations increase; it relies on fixed replenishment lead times and unified inventory cycles for ordering arrangements, without dynamically matching inventory support capacity with actual delivery cycles, which can easily lead to misalignment of support when delivery cycles change; and it relies on a hierarchical demand transmission model for inventory allocation decisions, lacking fine control over the dynamic balance between inventory holding costs and support risks, which can easily lead to difficulty in balancing cost constraints and operational stability. Summary of the Invention

[0005] To address this, the present invention provides an IoT platform supply chain management system based on multi-source data analysis. This system overcomes the problems in existing technologies, such as delayed demand response, low matching degree with changes in operational intensity, and slow matching speed, caused by over-reliance on historical periodic forecasts, by performing real-time quantitative calculations of actual consumption rhythms and operational load changes and synchronously correcting the forecast results.

[0006] To achieve the above objectives, this invention provides an IoT platform supply chain management system based on multi-source data analysis, comprising: an acquisition module, used to collect in real time the unit cycle consumption of key spare parts, vehicle operating intensity, existing inventory in the warehouse, in-transit transportation quantity, and average delivery time of corresponding suppliers for new energy buses in urban bus depots during operation; a consumption calculation module, used to calculate the predicted cycle consumption based on the unit cycle consumption, the vehicle operating intensity, and a preset intensity coefficient; a cycle calculation module, used to calculate the total inventory based on the existing inventory and the in-transit transportation quantity, and to calculate the inventory support cycle based on the total inventory and the predicted cycle consumption; and a status determination module, used to... The system compares the inventory support period with the average delivery time and determines whether the current situation is one of insufficient yard support based on the comparison result; a generation module calculates the replenishment lead time based on the average delivery time and the inventory support period, and generates a purchase requisition instruction or a transfer instruction between yards based on the replenishment lead time, the total inventory, and the predicted consumption period; a deviation calculation module obtains the actual spare parts consumption during the operation within a preset statistical period and calculates the consumption deviation value based on the actual spare parts consumption and the predicted consumption period; and a correction module corrects the preset intensity coefficient based on the deviation comparison result between the consumption deviation value and the preset deviation range.

[0007] Furthermore, the consumption calculation module includes: an intensity correction unit, which calculates an intensity correction consumption rate based on the unit cycle consumption and the vehicle operating intensity; a prediction calculation unit, which calculates a predicted cycle consumption based on a preset prediction cycle and the intensity correction consumption rate; and a prediction calculation unit, which calculates the predicted cycle consumption based on a historical fluctuation coefficient and the predicted cycle consumption, wherein the historical fluctuation coefficient is determined based on the average value and standard deviation of the predicted cycle consumption over a preset historical period.

[0008] Furthermore, the cycle calculation module includes: an inventory aggregation unit, which calculates the sum of the existing inventory quantity and the quantity in transit to obtain the total inventory; a daily consumption conversion unit, which calculates the ratio of the predicted cycle consumption to the preset prediction cycle to obtain the daily predicted consumption; and a cycle calculation unit, which calculates the ratio of the total inventory to the daily predicted consumption to obtain the inventory support cycle.

[0009] Furthermore, the status determination module includes: a comparison unit, which compares the inventory support period with the average delivery time to obtain the comparison result; and a status determination unit, which determines that the current status is one of insufficient site support when the comparison result indicates that the inventory support period is less than the average delivery time.

[0010] Further, the generation module includes: a replenishment calculation unit, which calculates the difference between the average delivery time and the inventory support period to obtain the replenishment lead time; an inventory status determination unit, which determines the inventory status based on the total inventory and the predicted consumption period; and a generation unit, which generates the purchase request instruction or the transfer instruction based on the observed duration fluctuation and the inventory status, wherein the observed duration fluctuation is determined based on the replenishment lead time within a preset generation period.

[0011] Furthermore, the inventory status determination unit includes: a redundancy calculation subunit, which is used to calculate the inventory redundancy rate based on the total inventory and the predicted periodic consumption; and a status determination subunit, which is used to determine that the inventory status is a pending allocation status when the inventory redundancy rate is greater than a preset redundancy threshold, and to determine that the inventory status is a low redundancy status when the inventory redundancy rate is less than or equal to the preset redundancy threshold.

[0012] Further, the generation unit includes: a duration statistics subunit, which is used to count all replenishment advance durations that are greater than a preset advance duration threshold within the preset generation cycle, and mark them as generation observation durations respectively; a duration distribution calculation subunit, which is used to calculate the standard deviation of all the generation observation durations to obtain the observation duration volatility; and a generation subunit, which is used to generate the purchase request instruction or the transfer instruction based on the comparison results of the observation duration volatility and the respective thresholds of the inventory status.

[0013] Furthermore, the generating subunit is used to determine that critical spare parts procurement is required when the observation duration fluctuation is less than or equal to a preset duration fluctuation threshold and the inventory status is the low redundancy status, so as to generate the procurement request instruction; and to determine that inter-site transfer is required when the observation duration fluctuation is greater than the preset duration fluctuation threshold and the inventory status is the pending allocation status, so as to generate the transfer instruction.

[0014] Furthermore, the deviation calculation module includes: an actual acquisition unit, which is used to acquire the actual spare parts consumption during the operation within a preset statistical period; and a deviation rate calculation unit, which is used to calculate the absolute value of the relative deviation between the sum of all actual spare parts consumption and the consumption in the predicted period, so as to obtain the consumption deviation value.

[0015] Furthermore, the correction module includes: an interval determination unit, which compares the consumption deviation value with the preset deviation interval to obtain the deviation comparison result; and a coefficient correction unit, which corrects the preset intensity coefficient based on the deviation comparison result and the consumption deviation value.

[0016] Compared with existing technologies, the beneficial effects of this invention are as follows: by weighting and normalizing multi-dimensional operational data such as vehicle mileage, runtime, and departure frequency to obtain operational intensity, and then dynamically correcting the unit cycle consumption based on operational intensity, the spare parts consumption prediction is transformed from being driven by a single historical average to an intensity-driven mode directly related to the actual vehicle operating load. Since spare parts wear is positively correlated with vehicle usage frequency, cumulative load, and start-stop frequency, an increase in operational intensity will simultaneously increase the wear rate per unit time, thereby enabling the predicted cycle consumption to adaptively adjust with changes in operating load. Simultaneously, the ratio of total inventory to predicted cycle consumption is used to calculate the inventory's ability to support a certain number of weeks. The system compares the actual consumption with the supplier's average delivery time, ensuring that replenishment decisions are based on a quantitative balance between consumption rate, inventory levels, and supply cycle. Furthermore, it uses the deviation between actual and predicted consumption to perform interval-based corrections on the intensity coefficient, enabling the system to form a closed-loop iterative mechanism based on operational feedback. This continuously narrows down prediction errors and avoids the risk of inventory backlog or supply disruptions caused by long-term accumulated deviations. The system can achieve dynamic and refined support and intelligent collaborative allocation of key spare parts for new energy bus depots, effectively solving the problems of delayed demand response, low matching degree with changes in operational intensity, and slow matching speed caused by over-reliance on historical periodic predictions when facing dynamic consumption scenarios.

[0017] Furthermore, by using unit cycle consumption as the basic consumption rate and introducing vehicle operating intensity to proportionally amplify it, the consumption rate increases synchronously with the increase in operating load, reflecting the positive correlation between operating load and spare parts wear. Then, through preset intensity coefficients for overall calibration, the model can absorb systematic influencing factors such as differences in the site environment and vehicle technical condition, achieving unified adjustment of basic consumption. Based on this, the prediction period is used for linear expansion in the time dimension, making the estimated cycle consumption increase proportionally with the continuous operating time, reflecting the cumulative effect. Furthermore, a volatility coefficient is constructed using standard deviation and mean, transforming historical dispersion into a risk correction ratio for future demand. This allows the predicted cycle consumption to maintain a trend-growth logic while possessing elastic response capabilities to uncertainty, thereby achieving synergistic coupling between consumption trends, operating load, and historical volatility. This improves prediction accuracy and the reliability of inventory decisions, ultimately enhancing the dynamic control capabilities of supply chain management under the IoT platform.

[0018] Furthermore, by superimposing existing inventory with the quantity in transit to obtain the total inventory, the system comprehensively depicts the resources that can be actually controlled in the future in terms of quantity. Then, the predicted consumption is normalized according to a preset prediction period, converting it into an average daily predicted consumption, establishing a one-to-one correspondence between consumption rate and time scale. Finally, the inventory sustainability period is calculated by the ratio of total inventory to average daily predicted consumption, establishing a direct functional relationship between inventory and consumption rate per unit time. Since the inventory sustainability period is essentially equal to the resource inventory divided by the consumption intensity per unit time, the sustainability period naturally shortens when the consumption rate increases and extends accordingly when inventory increases. A continuous mapping chain of quantity, time, and rate is formed among the parameters, avoiding dimensional confusion and estimation amplification errors. This allows the judgment of inventory assurance capability to be based on quantifiable and derivable mathematical relationships, achieving accurate assessment of the supply chain assurance period.

[0019] Furthermore, by directly comparing the inventory sustainability period with the average delivery time, a unified benchmark is established between the time scale corresponding to inventory depletion and the time required for supply replenishment: when the inventory sustainability period is less than the average delivery time, it means that the inventory will be exhausted before replenishment is completed; when the inventory sustainability period is greater than or equal to the average delivery time, it means that the inventory can cover the replenishment waiting time. Under this judgment logic, a time-dimensional mapping is formed between the inventory sustainability period and the supply arrival time. The higher the consumption rate or the lower the inventory, the shorter the inventory sustainability period, and the higher the probability of triggering a state of insufficient guarantee; conversely, the system maintains a safe range, thereby transforming the inventory safety judgment into a quantifiable time boundary problem, avoiding the lag risk brought by empirical judgment, and realizing timely and objective identification of the site's guarantee status.

[0020] Furthermore, by converting the difference between the average delivery time and the inventory's sustainable cycle into replenishment lead time, the time difference between inventory consumption time and supply arrival time is explicitly quantified. A larger time difference indicates a higher risk of future supply disruptions and a stronger urgency for replenishment. Simultaneously, the inventory structure is characterized by the ratio between total inventory and the predicted consumption rate, creating a corresponding mapping between inventory size and consumption intensity. Higher total inventory and lower consumption intensity indicate a more redundant inventory state, while a smaller inventory state indicates a more strained state. Furthermore, by statistically analyzing the dispersion of replenishment lead time within the generation cycle, the volatility of the observed time is obtained, incorporating the stability and volatility risk of the time difference into the decision-making process. When the time difference fluctuates significantly, priority is given to resource allocation between stations to balance resources. When the time difference is stable and the inventory is in a low-redundancy state, a purchase request is triggered. Thus, replenishment decisions are simultaneously constrained by the size of the time difference, the inventory structure ratio, and the time volatility characteristics, achieving a linkage and matching between supply and demand rhythms and inventory dynamics.

[0021] Furthermore, by constructing an inventory redundancy rate from the total inventory and the predicted consumption over the period, a proportional relationship is established between inventory size and future consumption demand. The larger the total inventory relative to the predicted consumption, the higher the redundancy rate, indicating a more ample capacity to support future consumption within a given timeframe. Conversely, when the total inventory is close to or below the predicted consumption, the redundancy rate tends to decrease, indicating a weakening ability of inventory to cover future consumption. Further, by pre-setting redundancy thresholds to divide the redundancy rate into intervals, the determination of inventory status is based on a quantitative comparison between inventory supply capacity and demand intensity. This transforms whether inventory is surplus or shortage into a quantifiable proportional judgment logic, allowing changes in inventory size to directly reflect status switching results. This provides a clear basis for subsequent allocation or procurement decisions, thereby improving the targeting and dynamic response capability of inventory resource allocation.

[0022] Furthermore, by thresholding the replenishment lead time and generating observation durations, and then quantifying their dispersion using standard deviation, the concentration or dispersion of time differences can objectively reflect the matching degree between demand rhythm and inventory consumption rate. When the fluctuation of the observation duration is low, it indicates that the consumption trend is relatively stable, and combined with the inventory status threshold, a purchase request can be directly triggered to fill the gap. When the fluctuation is high and the inventory is in a pending allocation state, it indicates that the lead time changes significantly but there are still adjustable resources within the system. Through inter-site allocation, inventory structure optimization can be achieved without increasing external procurement costs. By coupling the time difference distribution characteristics with inventory structure indicators, the decision-making logic is upgraded from a single numerical judgment to the coordinated control of fluctuation characteristics and inventory status.

[0023] Furthermore, by combining the fluctuation of observation duration with inventory status, the decision-making logic simultaneously considers the stability of demand changes and the redundancy level of the inventory structure: when the dispersion of replenishment lead time is small, it indicates that the consumption trend is relatively stable. If the inventory is in a low redundancy state, the future gap is persistent and predictable, making it suitable to replenish through procurement to form a stable supply. However, when the fluctuation of replenishment lead time is large, it reflects that demand changes have phased fluctuations or uncertainties. If the inventory is redundant, structural rebalancing is prioritized through inter-site transfers, which can alleviate local tensions without increasing the overall inventory volume. This allows the inventory flow to match the consumption fluctuation trend, achieving dynamic optimization of supply chain resources.

[0024] Furthermore, by classifying the outbound quantity within a preset statistical period by time, deducting the returned quantity in reverse, and correcting cross-period data boundaries, the actual spare parts consumption truly reflects the net outflow of materials from inventory to the operating system within that period. Based on this, the absolute value of the relative deviation between this total net consumption and the predicted consumption is calculated, allowing the deviation to simultaneously reflect the directional elimination and amplitude amplification characteristics of demand forecasting errors. This enables a quantitative comparison of inventory outflow intensity and predicted intensity on the same time scale. Since consumption is essentially directly related to vehicle operating load and inventory release processes, this calculation method establishes a one-to-one numerical mapping relationship between changes in operating load and changes in inventory quantity, providing a clear quantitative basis for subsequent intensity coefficient corrections, improving the system's self-calibration capability and operational stability, and ultimately achieving dynamic closed-loop control of supply chain management.

[0025] Furthermore, by dividing consumption deviation values ​​into three ranges—underestimation, allowable, and overestimation—the quantitative difference between the forecast and actual net consumption is clearly quantified and graded. When actual consumption is lower than the forecast, the preset intensity coefficient is reduced proportionally to the deviation, causing subsequent forecast consumption to decrease accordingly, thus avoiding capital tied up due to persistently overestimated inventory. When actual consumption is higher than the forecast, the preset intensity coefficient is increased proportionally to the deviation, causing the forecast consumption to be adjusted upwards to compensate for the impact of increased operational intensity on spare parts consumption rates. When the deviation is within the allowable range, the coefficient remains unchanged to prevent the system from over-adjusting within normal fluctuation ranges. This correction method ensures that the change in intensity coefficient corresponds proportionally to the actual deviation, enabling the forecast model to automatically converge and stabilize with changes in operating load, ultimately achieving adaptive dynamic correction of supply chain consumption forecasts. Attached Figure Description

[0026] Figure 1 is a schematic diagram of the IoT platform supply chain management system based on multi-source data analysis in this embodiment; Figure 2 is a logic diagram of the determination of insufficient support status of the site by the status determination unit in this embodiment; Figure 3 is a logic diagram of the determination of inventory status by the status determination subunit in this embodiment; Figure 4 is a logic diagram of the determination of instructions generated by the generation subunit in this embodiment. Detailed Implementation

[0027] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0028] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0029] Please refer to Figure 1, which is a schematic diagram of the IoT platform supply chain management system based on multi-source data analysis in this embodiment. This embodiment provides an IoT platform supply chain management system based on multi-source data analysis, including: an acquisition module, which is used to collect in real time the unit cycle consumption of each key spare part, vehicle operation intensity, existing inventory quantity in the warehouse, in-transit transportation quantity, and the average delivery time of the corresponding supplier for new energy buses in urban bus stations during operation; a consumption calculation module, which is connected to the acquisition module, and is used to calculate the predicted cycle consumption based on the unit cycle consumption, the vehicle operation intensity, and a preset intensity coefficient; a cycle calculation module, which is connected to both the consumption calculation module and the acquisition module, and is used to calculate the total inventory based on the existing inventory quantity and the in-transit transportation quantity, and calculate the inventory support cycle based on the total inventory and the predicted cycle consumption; and a status determination module, which is connected to both the cycle calculation module and the acquisition module. The system is connected to the acquisition module to compare the inventory support period with the average delivery time and determine whether the current situation is one of insufficient yard support based on the comparison result. The generation module, connected to both the status determination module and the acquisition module, calculates the replenishment lead time based on the insufficient yard support status, the average delivery time, and the inventory support period, and generates a purchase request instruction or a yard transfer instruction based on the replenishment lead time, the total inventory, and the predicted period consumption. The deviation calculation module, connected to both the consumption calculation module and the generation module, acquires the actual spare parts consumption during the operation within a preset statistical period and calculates the consumption deviation value based on the actual spare parts consumption and the predicted period consumption. The correction module, connected to both the deviation calculation module and the consumption calculation module, corrects the preset intensity coefficient based on the deviation comparison result between the consumption deviation value and the preset deviation range.

[0030] In this embodiment, the acquisition module includes: a data acquisition unit, which is used to collect in real time the number of key spare parts issued from the warehouse, as well as the operating mileage, running time and departure frequency of new energy buses; and an intensity calculation unit, which is used to calculate the vehicle operating intensity based on the operating mileage, running time and departure frequency.

[0031] In this embodiment, critical spare parts refer to components that perform core functions during the operation of new energy buses, have high fault sensitivity, and whose failure will directly affect the safe operation, attendance rate, or operational continuity of the vehicle. These typically include, but are not limited to, power battery-related components such as battery module connectors and thermal management unit components, electric drive system components such as motor controller power modules and bearings, brake system wear parts such as brake pads and brake valve assemblies, key steering and suspension connectors, and high-voltage electrical connectors. These spare parts generally have relatively clear consumption cycles, are closely related to the intensity of vehicle operation, have relatively long replenishment cycles, or have high unit prices. Therefore, they require key monitoring and dynamic predictive management.

[0032] In this embodiment, the data acquisition unit achieves data acquisition by interfacing with the warehouse management system and vehicle dispatch monitoring system within the bus depot. Specifically, the outbound quantity of each key spare part is obtained from the outbound record data of the warehouse management system. This outbound record data is entered by the warehouse management terminal using a barcode scanner or RFID identification device when spare parts are issued, and includes fields for spare part number, outbound time, and outbound quantity. The acquisition unit reads and summarizes the outbound quantity within the corresponding time period according to a preset acquisition cycle. The operating mileage and running time of new energy buses are collected through the vehicle's onboard terminal. The onboard terminal is connected to the vehicle control system or odometer signal line, recording the vehicle's cumulative mileage and start-up time data in real time, and uploading it to the vehicle dispatch monitoring system via 4G / 5G or wired network. The departure schedule is obtained from the departure record data generated by the dispatch system. The departure record data includes the vehicle number and departure time. The acquisition unit counts and statistically analyzes the departure records according to a preset statistical cycle to obtain the corresponding vehicle's departure schedule, i.e., the number of departures per unit time.

[0033] The preset intensity coefficient is a proportional parameter used to adjust the coupling relationship between vehicle operating intensity and unit cycle consumption. Its setting is mainly based on: the average deviation between actual consumption and theoretical consumption calculated according to the basic model over multiple past statistical periods; the elasticity coefficient of intensity changes on consumption increase under different operating conditions, such as peak operation, off-peak operation, and extreme weather; the fatigue characteristics and wear acceleration range of key spare parts materials; and the fluctuation distribution of the vehicle operating intensity index. Typically, the value range is determined by regression analysis or proportional calibration of data from at least six consecutive statistical periods, generally set between 0.9 and 1.3. In this embodiment, it is set to 1.12, which can moderately amplify predicted consumption when operating intensity is high and suppress over-prediction when operating intensity is low, keeping the prediction results stable under high and low operating condition switching scenarios.

[0034] The preset statistical period is a time window used to summarize actual spare parts consumption data and verify prediction deviations. Its setting is mainly based on: the average replacement cycle and minimum replacement interval of key spare parts; the number of days of inventory safety reserve and replenishment lead time; the typical fluctuation cycle of vehicle operating intensity, such as weekly or bi-weekly cycles; and the variance stability interval of historical consumption data. Typically, a time length is selected that covers at least one complete operational intensity fluctuation cycle and has a sample size of no less than 30 valid consumption records, generally set between 10 and 21 days. In this embodiment, it is set to 14 days, which ensures the representativeness of the statistical results while avoiding correction lag due to an excessively long period, thereby improving the timeliness of dynamic parameter calibration.

[0035] In this embodiment, the acquisition module also includes a consumption calculation unit. The process of calculating the unit cycle consumption C based on replacement record data, inbound quantity, and outbound quantity includes: firstly, acquiring the outbound quantity recorded by the warehousing system within a unit collection time t, and combining it with the spare parts replacement record data in the vehicle maintenance system to match and verify the outbound quantity, so as to eliminate non-actual consumption data caused by allocation, inventory adjustment, or abnormal outbound, thereby determining the effective outbound quantity N; at the same time, by comparing the inbound quantity and inventory change within the same statistical period, the integrity of the outbound quantity is verified to avoid statistical deviations caused by delayed accounting or inventory adjustment; after obtaining the effective outbound quantity N, the unit cycle consumption C is calculated according to the following formula: C=N / t; it can be understood that since the replacement of spare parts for new energy buses usually adopts the maintenance mode of immediate installation upon receipt, once the spare parts are out of the warehouse, they enter the repair or replacement process, so the outbound quantity can directly reflect the actual consumption.

[0036] The process of calculating the vehicle operation intensity K based on the operating mileage, operating time, and departure frequency includes: K=α×L / L0+β×H / H0+γ×F / F0, where α is the preset mileage weight, L is the operating mileage, L0 is the preset standard reference mileage, β is the preset time weight, H is the operating time, H0 is the preset standard reference time, γ is the preset departure frequency weight, F is the departure frequency, and F0 is the preset standard reference frequency.

[0037] The preset mileage weight is a proportional parameter used to characterize the contribution of operating mileage to spare parts wear. It depends on the wear sensitivity of the target critical spare parts under unit mileage conditions, such as the load difference under different line conditions such as straight roads, sloping road sections, and congestion levels, as well as the correlation coefficient analysis results between historical mileage and replacement frequency. It is usually set between 0.3 and 0.6. In this embodiment, it is set to 0.45, which can reflect the dominant influence of mileage on structural wear spare parts, while avoiding the masking of other operating factors.

[0038] The preset standard reference mileage is a benchmark value used to normalize the actual operating mileage. It depends on the statistical average of the daily operating mileage of new energy buses in the station under stable operating conditions, the median of the operating data of the past year, and the sample range of the stable period of the route structure. It is usually set between 150 km and 250 km. In this embodiment, it is set to 200 km, which can make the operating intensity index close to the benchmark level under normal operating conditions.

[0039] The preset duration weight is a proportional parameter used to measure the impact of running time on the aging or thermal decay of spare parts. It depends on the thermal accumulation effect of key spare parts under continuous working conditions, the proportion of idling time, and the regression slope between historical running time and failure rate. It is usually set between 0.2 and 0.4. In this embodiment, it is set to 0.3, which can reflect the impact of long-term operation on electrical and rubber seal spare parts.

[0040] The preset standard reference duration is a reference benchmark for standardizing the actual running time. It depends on the average daily online operating time of vehicles, the scheduling structure, and the distribution characteristics of peak and off-peak periods. It is usually set between 8 and 14 hours. In this embodiment, it is set to 10 hours to ensure that the intensity calculation maintains a stable benchmark under the regular shift arrangement.

[0041] The preset shift weight is a proportional parameter used to measure the impact of changes in departure frequency on the start-stop impact of spare parts. It depends on the impact of the number of start-stops on mechanically impacted components, the impact of frequent departures on the load fluctuation of the control system, and the correlation analysis results of historical shifts and maintenance records. It is usually set between 0.1 and 0.3. In this embodiment, it is set to 0.25, which can reflect the additional consumption effect of high-frequency start-stops on vulnerable components.

[0042] The preset standard reference number is a benchmark value that normalizes the actual number of departures. It depends on the average number of departures per day, the line density distribution, and the annual scheduling statistics of the depot under normal dispatching mode. It is usually set between 40 and 80 departures. In this embodiment, it is set to 60 departures, which can keep the operating intensity within the benchmark range under standard dispatching conditions.

[0043] By weighted and normalized multi-dimensional operational data such as vehicle mileage, runtime, and departure frequency, the operational intensity is obtained. This operational intensity is then used to dynamically adjust the unit cycle consumption, transforming spare parts consumption forecasting from a single historical average-driven model to an intensity-driven model directly related to the actual vehicle operating load. Since spare parts wear is positively correlated with vehicle usage frequency, cumulative load, and start-stop frequency, increased operational intensity will simultaneously increase the wear rate per unit time, thus adaptively adjusting the predicted cycle consumption according to changes in operating load. Simultaneously, the inventory support period is calculated by the ratio of total inventory to the predicted cycle consumption, and this ratio is compared with the average delivery rate from suppliers. By comparing the consumption time, replenishment decisions are based on a quantitative balance between consumption rate, inventory level, and supply cycle. Furthermore, the intensity coefficient is adjusted in intervals using the deviation between actual and predicted consumption, enabling the system to form a closed-loop iterative mechanism based on operational feedback. This continuously narrows down prediction errors and avoids the risk of inventory backlog or supply disruption caused by long-term accumulated deviations. It can achieve dynamic and refined support and intelligent collaborative allocation of key spare parts for new energy bus depots, effectively solving the problems of lagging demand response, low matching degree with changes in operational intensity, and slow matching speed when facing dynamic consumption scenarios due to over-reliance on historical periodic predictions.

[0044] Specifically, the consumption calculation module includes: an intensity correction unit, which calculates an intensity correction consumption rate based on the unit cycle consumption and the vehicle operating intensity, R=C×K×k, where R is the intensity correction consumption rate and k is a preset intensity coefficient; a prediction calculation unit, which calculates the predicted cycle consumption based on a preset prediction period and the intensity correction consumption rate, Q1=R×T, where Q1 is the predicted cycle consumption and T is the preset prediction period; and a prediction calculation unit, which calculates the predicted cycle consumption based on a historical fluctuation coefficient and the predicted cycle consumption, where the historical fluctuation coefficient is determined based on the average and standard deviation of the predicted cycle consumption over a preset historical period, Q2=Q1×(1+δ), δ=σ / μ, where Q2 is the predicted cycle consumption, δ is the historical fluctuation coefficient, σ is the standard deviation of the predicted cycle consumption over a preset historical period, and μ is the average of the predicted cycle consumption over a preset historical period.

[0045] The preset forecast period refers to the target time span used to calculate future consumption. It depends on the distribution characteristics of the supplier's average delivery time, the logic of setting the inventory safety factor, the time superposition relationship between approval flow and transportation links, and the linear amplification law of consumption rate with time scale. Its setting must ensure that the forecast period is not less than the time required for a complete procurement response chain and is consistent with the calculation scale of the inventory's support cycle to avoid decision distortion caused by time dimension mismatch; at the same time, it must be controlled within the range where errors can accumulate to prevent the uncertainty caused by excessively long forecast periods from amplifying intensity fluctuations. It is typically set between 7 and 30 days; in this embodiment, it is set to 14 days, which can cover a replenishment response cycle while forming a closed loop with the forecast results and inventory turnover rhythm on a consistent time scale.

[0046] The preset historical duration refers to the historical sample time range selected for statistical analysis of historical fluctuation coefficients. It depends on the periodic variation pattern of vehicle operating intensity, the time stability interval of spare parts failure distribution, the impact of sample size on the statistical stability of standard deviation and mean, and the degree of interference from structural changes on data representativeness. Its setting must ensure that the sample size is sufficient to stabilize σ / μ, while covering at least one complete operating load change cycle to avoid coefficient distortion caused by occasional fluctuations; it must also avoid introducing structural factors such as technology upgrades and route adjustments due to an excessively long time span. It is typically set between 3 and 12 months; in this embodiment, it is set to 6 months, which ensures statistical stability while allowing the historical fluctuation coefficients to accurately reflect the consumption dispersion of the current operating phase.

[0047] By using unit cycle consumption as the basic consumption rate and incorporating vehicle operating intensity to proportionally amplify it, the consumption rate increases synchronously with the increase in operating load, reflecting the positive correlation between operating load and spare parts wear. Furthermore, through pre-set intensity coefficients for overall calibration, the model can absorb systematic influencing factors such as differences in the depot environment and vehicle technical condition, achieving unified adjustment of basic consumption. On this basis, the prediction period is used for linear expansion in the time dimension, making the estimated cycle consumption increase proportionally with the continuous operating time, reflecting the cumulative effect. Further, a volatility coefficient is constructed using standard deviation and mean, transforming historical dispersion into a risk correction ratio for future demand. This allows the predicted cycle consumption to maintain a trend-growth logic while possessing elastic response capabilities to uncertainty, thereby achieving synergistic coupling between consumption trends, operating load, and historical volatility. This improves prediction accuracy and the reliability of inventory decisions, ultimately enhancing the dynamic control capabilities of supply chain management under the IoT platform.

[0048] Specifically, the cycle calculation module includes: an inventory aggregation unit, which calculates the sum of the existing inventory quantity and the quantity in transit to obtain the total inventory; a daily consumption conversion unit, which calculates the ratio of the predicted cycle consumption to the preset prediction cycle to obtain the daily predicted consumption; and a cycle calculation unit, which calculates the ratio of the total inventory to the daily predicted consumption to obtain the inventory support cycle.

[0049] By overlaying existing inventory with in-transit shipments to obtain the total inventory, the system comprehensively depicts future available resources in terms of quantity. Then, the predicted consumption is normalized to a pre-defined forecast period, converting it into an average daily predicted consumption, establishing a one-to-one correspondence between consumption rate and time scale. Finally, the inventory sustainability period is calculated by the ratio of total inventory to average daily predicted consumption, establishing a direct functional relationship between inventory and consumption rate per unit time. Since the inventory sustainability period is essentially equal to the resource inventory divided by the consumption intensity per unit time, the sustainability period naturally shortens as the consumption rate increases and lengthens accordingly as inventory increases. This forms a continuous mapping chain of quantity, time, and rate among the parameters, avoiding dimensional confusion and amplified estimation errors. Therefore, the assessment of inventory assurance capability is based on quantifiable and derivable mathematical relationships, achieving accurate evaluation of supply chain assurance periods.

[0050] Please refer to Figure 2, which is a logic diagram for determining the insufficient support status of the site by the status determination unit in this embodiment. The status determination module includes: a comparison unit, which is used to compare the inventory support period with the average delivery time to obtain the comparison result; and a status determination unit, which is used to determine that the site is currently in a state of insufficient support when the comparison result is that the inventory support period is less than the average delivery time.

[0051] By directly comparing the inventory sustainability period with the average delivery time, a unified benchmark is established between the time scale corresponding to inventory depletion and the time required for supply replenishment: when the inventory sustainability period is less than the average delivery time, it means that the inventory will be exhausted before replenishment is completed; when the inventory sustainability period is greater than or equal to the average delivery time, it means that the inventory can cover the replenishment waiting time. Under this judgment logic, a time-dimensional mapping is formed between the inventory sustainability period and the supply arrival time. The higher the consumption rate or the lower the inventory, the shorter the inventory sustainability period, and the higher the probability of triggering a state of insufficient guarantee; conversely, the system maintains a safe range, thereby transforming the inventory safety judgment into a quantifiable time boundary problem, avoiding the lag risk brought by empirical judgment, and achieving timely and objective identification of the site's guarantee status.

[0052] Specifically, the generation module includes: a replenishment calculation unit, which calculates the difference between the average delivery time and the inventory support period to obtain the replenishment lead time; an inventory status determination unit, which determines the inventory status based on the total inventory and the predicted period consumption; and a generation unit, which generates the purchase request instruction or the transfer instruction based on the observed duration fluctuation and the inventory status, wherein the observed duration fluctuation is determined based on the replenishment lead time within a preset generation period.

[0053] By converting the difference between average delivery time and the inventory's sustainable cycle into replenishment lead time, the time difference between inventory consumption time and supply arrival time is explicitly quantified. A larger time difference indicates a higher risk of future supply disruptions and a stronger urgency for replenishment. Simultaneously, the inventory structure is characterized by the ratio between total inventory and predicted consumption over the cycle, creating a corresponding mapping between inventory size and consumption intensity. Higher total inventory and lower consumption intensity indicate a more redundant inventory state, while a smaller inventory state indicates a more strained state. Furthermore, by statistically analyzing the dispersion of replenishment lead time within the generation cycle, the volatility of the observed time is obtained, incorporating the stability and volatility risk of the time difference into the decision-making process. When the time difference fluctuates significantly, inter-site allocation is prioritized to balance resources. When the time difference is stable and inventory is in a low-redundancy state, a purchase request is triggered. Thus, replenishment decisions are simultaneously constrained by the size of the time difference, the inventory structure ratio, and the time volatility characteristics, achieving a linkage and matching between supply and demand rhythms and inventory dynamics.

[0054] Please refer to Figure 3, which is a logic diagram for determining the inventory status by the status determination subunit in this embodiment. In this embodiment, the inventory status determination unit includes: a redundancy calculation subunit, which is used to calculate the inventory redundancy rate based on the total inventory and the predicted periodic consumption; and a status determination subunit, which is used to determine that the inventory status is a pending allocation status when the inventory redundancy rate is greater than a preset redundancy threshold, and to determine that the inventory status is a low redundancy status when the inventory redundancy rate is less than or equal to the preset redundancy threshold.

[0055] The preset redundancy threshold is a proportional boundary value used to divide inventory into a surplus or low redundancy state. It depends on multi-dimensional operational constraints such as the statistical dispersion of the consumption rate of key spare parts, the average delivery time of suppliers and its fluctuation range, the distribution characteristics of historical inventory turnover cycles, the maximum tolerable downtime of the depot, and the upper limit of capital occupation. In specific settings, interval stratification analysis can be performed based on the distribution of the difference between the inventory support cycle and the actual delivery cycle over several past statistical periods. The risk buffer ratio is determined by combining the mean and standard deviation of the consumption deviation value, so that the threshold can cover the safety boundary of the high volatility range without exceeding the economic range of inventory turnover efficiency. It is usually set between 0.1 and 0.5. In this embodiment, it is set to 0.3, which can form a quantifiable risk buffer while maintaining inventory turnover efficiency, thereby improving the stability and executability of inventory status determination.

[0056] By constructing an inventory redundancy rate from the total inventory and the predicted consumption over the period, a proportional relationship is established between inventory size and future consumption demand. The larger the total inventory relative to the predicted consumption, the higher the redundancy rate, indicating a more ample capacity to support future consumption within a given timeframe. Conversely, when the total inventory is close to or below the predicted consumption, the redundancy rate tends to decrease, indicating a weakening ability of the inventory to cover future consumption. Furthermore, by pre-setting redundancy thresholds to divide the redundancy rate into intervals, the determination of inventory status is based on a quantitative comparison between inventory supply capacity and demand intensity. This transforms whether inventory is surplus or shortage into a quantifiable proportional judgment logic, allowing changes in inventory size to directly reflect status transitions. This provides a clear basis for subsequent allocation or procurement decisions, thereby improving the targeting and dynamic responsiveness of inventory resource allocation.

[0057] Specifically, the generation unit includes: a duration statistics subunit, which is used to count all replenishment advance durations that are greater than a preset advance duration threshold within the preset generation cycle, and mark them as generation observation durations respectively; a duration distribution calculation subunit, which is used to calculate the standard deviation of all the generation observation durations to obtain the observation duration volatility; and a generation subunit, which is used to generate the purchase request instruction or the transfer instruction based on the comparison results of the observation duration volatility and the respective thresholds of the inventory status.

[0058] The preset lead time threshold is a benchmark value used to screen replenishment lead times with practical regulatory significance. It depends on a comprehensive evaluation of multiple factors, including the standard deviation distribution of the supplier's historical delivery cycle, the timeliness fluctuation range of the transportation link, the shortest achievable cycle of key spare parts from order placement to warehousing, the number of days of safety stock coverage at the depot, and the fault tolerance space of the vehicle maintenance plan. Its setting needs to be determined in conjunction with the intersection range of the delivery cycle probability distribution curve and the inventory consumption slope to avoid misjudging short-term statistical disturbances as real shortage risks. It is usually set between 1.5 days and 3 days. In this embodiment, it is set to 2 days, which can suppress excessively frequent triggering of replenishment judgments while ensuring early response capability, thereby improving the stability and accuracy of instruction generation logic.

[0059] By thresholding replenishment lead time and generating observation durations, and then quantifying their dispersion using standard deviation, the concentration or dispersion of time differences can objectively reflect the matching degree between demand rhythm and inventory consumption rate. When the fluctuation of the observation duration is low, it indicates a relatively stable consumption trend, and combined with inventory status thresholds, a purchase request can be directly triggered to fill the gap. When the fluctuation is high and the inventory is in a pending allocation state, it indicates that the lead time has changed significantly, but there are still adjustable resources within the system. Through inter-site allocation, inventory structure optimization can be achieved without increasing external procurement costs. By coupling the time difference distribution characteristics with inventory structure indicators, the decision-making logic is upgraded from a single numerical judgment to the coordinated control of fluctuation characteristics and inventory status.

[0060] Please refer to Figure 4, which is a logic diagram for determining the generation instruction of the generation subunit in this embodiment. In this embodiment, the generation subunit is used to determine that critical spare parts need to be procured when the observation duration fluctuation is less than or equal to a preset duration fluctuation threshold and the inventory status is the low redundancy status, so as to generate the procurement application instruction; and to determine that inter-site transfer needs to be carried out when the observation duration fluctuation is greater than the preset duration fluctuation threshold and the inventory status is the pending allocation status, so as to generate the transfer instruction.

[0061] The preset lead time fluctuation threshold is used to determine whether the dispersion of replenishment lead time exceeds the normal operating range. Its setting is based on the following: First, statistical analysis is performed on replenishment lead time data from at least 12 past generation cycles to calculate the mean and standard deviation; second, the coefficient of variation is calculated, and the actual proportion of emergency purchases or inter-site transfers occurring within different coefficient of variation ranges is analyzed; when the coefficient of variation exceeds a certain value, the frequency of abnormal transfers increases significantly and the inventory redundancy utilization rate increases by more than 10%, then this inflection point of the coefficient of variation is used as the threshold reference value; it is usually set between 0.08 and 0.25, and in this embodiment, it is set to 0.12, which can promptly identify stages of amplified demand fluctuations while maintaining stable inventory turnover, thus improving the accuracy of purchase and transfer judgments.

[0062] In this embodiment, when the observation duration fluctuation is less than or equal to the preset duration fluctuation threshold and the inventory status is pending allocation, it indicates that the demand fluctuation is relatively stable but there is a phased redundancy in the local inventory structure. At this time, the existing inventory structure can be maintained first, and only the inventory distribution is marked and tracked without immediately triggering allocation or procurement operations. When the observation duration fluctuation is greater than the preset duration fluctuation threshold and the inventory status is low redundancy, it indicates that the demand fluctuates uncertainly but the overall inventory safety margin is insufficient. At this time, the generation subunit can enter the early warning observation state, shorten the generation cycle and increase the data refresh frequency so that it can re-determine whether to trigger procurement or allocation instructions in the next generation cycle, thereby avoiding over-response caused by short-term abnormal fluctuations.

[0063] By combining the fluctuation of observation time with inventory status, the decision-making logic considers both the stability of demand changes and the redundancy level of inventory structure: when the dispersion of replenishment lead time is small, it indicates that the consumption trend is relatively stable. If the inventory is in a low redundancy state, the future gap is persistent and predictable, and it is suitable to replenish through procurement to form a stable supply. However, when the fluctuation of replenishment lead time is large, it reflects that the demand changes have phased fluctuations or uncertainties. If the inventory is redundant, structural rebalancing is prioritized through inter-site transfers. This can alleviate local tensions without increasing the total inventory, thereby matching the inventory flow with the consumption fluctuation trend and achieving dynamic optimization of supply chain resources.

[0064] Specifically, the deviation calculation module includes: an actual acquisition unit, which is used to acquire the actual spare parts consumption during the operation within a preset statistical period; and a deviation rate calculation unit, which is used to calculate the absolute value of the relative deviation between the sum of all actual spare parts consumption and the consumption in the predicted period, so as to obtain the consumption deviation value.

[0065] In this embodiment, the actual acquisition unit interfaces with the warehouse management system within the urban public transport station to read the outbound records of key spare parts in real time within a preset statistical period. The outbound record data includes the spare part number, outbound time, outbound quantity, and corresponding vehicle number. The actual acquisition unit filters and sums up all outbound quantities within the preset statistical period from the start time to the end time to obtain the actual spare parts consumption within the preset statistical period. When there are return records, the return quantity is deducted in reverse. When spare parts are outbound across periods, they are assigned according to the statistical period to which the outbound time belongs, thereby ensuring that the actual spare parts consumption is the net consumption quantity within the preset statistical period.

[0066] By classifying the outbound quantity within a preset statistical period by time, deducting the returned quantity in reverse, and correcting cross-period data boundaries, the actual spare parts consumption truly reflects the net outflow of materials from inventory to the operating system within that period. Based on this, the absolute value of the relative deviation between this net consumption and the predicted consumption is calculated. This deviation value simultaneously reflects the directional elimination and amplitude amplification characteristics of demand forecasting errors, thus quantifying and comparing the intensity of inventory outflow with the predicted intensity on the same time scale. Since consumption is essentially directly related to vehicle operating load and inventory release processes, this calculation method establishes a one-to-one numerical mapping relationship between changes in operating load and changes in inventory quantity. This provides a clear quantitative basis for subsequent intensity coefficient corrections, enhances the system's self-calibration capability and operational stability, and ultimately achieves dynamic closed-loop control of supply chain management.

[0067] Specifically, the correction module includes: an interval determination unit, which compares the consumption deviation value with the preset deviation interval to obtain the deviation comparison result; and a coefficient correction unit, which corrects the preset intensity coefficient based on the deviation comparison result and the consumption deviation value.

[0068] In this embodiment, the interval determination unit is used to determine the deviation comparison result as an underestimation of consumption when the consumption deviation value is less than the minimum value of the preset deviation interval, and to determine the deviation comparison result as an allowable deviation when the consumption deviation value is within the preset deviation interval, and to determine the deviation comparison result as an overestimation of consumption when the consumption deviation value is greater than the maximum value of the preset deviation interval.

[0069] The coefficient correction unit is used to reduce the preset intensity coefficient, k'=k×(1-D), when the judgment deviation comparison result is an underestimation of consumption, where k' is the reduced preset intensity coefficient and D is the consumption deviation value; and when the judgment deviation comparison result is an allowable deviation, no correction is required; and when the judgment deviation comparison result is an overestimation of consumption, the preset intensity coefficient, k'=k×(1+D), is increased.

[0070] The preset deviation range is a positive range starting from zero, where the consumption deviation is the absolute value of the relative deviation and is always positive. Its upper limit depends on the quantile interval of the statistical distribution of the absolute deviation value within at least 12 consecutive statistical periods, the stable segment of the proportion of the deviation standard deviation to the average value, and the inflection point of the increase in inventory capital occupancy rate and vehicle downtime probability due to missing parts at different error levels. When the absolute deviation exceeds the upper limit, the inventory turnover efficiency or guarantee risk shows a significant amplification trend, and the corresponding range is used as the control range. Usually, the range is set between [0, 20%], and in this embodiment, it is set to [0, 10%], which can cover the fluctuations of normal operation and suppress the cumulative expansion of error, ensuring the stability and controllability of the prediction correction mechanism.

[0071] By categorizing consumption deviations into three ranges—underestimation, allowable, and overestimation—the quantitative difference between the forecast and actual net consumption is clearly quantified and graded. When actual consumption is lower than the forecast, a preset intensity coefficient is reduced proportionally to the deviation, causing subsequent forecast consumption to decrease accordingly, thus avoiding capital tied up due to persistently overestimated inventory. When actual consumption is higher than the forecast, the preset intensity coefficient is increased proportionally to the deviation, causing the forecast consumption to be adjusted upwards to compensate for the impact of increased operational intensity on spare parts consumption rates. When the deviation is within the allowable range, the coefficient remains unchanged to prevent the system from over-adjusting within normal fluctuations. This correction method maintains a proportional relationship between the magnitude of the intensity coefficient change and the magnitude of the actual deviation, enabling the forecasting model to automatically converge and stabilize with changes in operating load, ultimately achieving adaptive dynamic correction of supply chain consumption forecasts.

[0072] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A supply chain management system for an IoT platform based on multi-source data analysis, characterized in that, include: The acquisition module is used to collect data in real time on the unit cycle consumption of key spare parts, vehicle operation intensity, warehouse inventory, transportation quantity, and average delivery time of corresponding suppliers for new energy buses in urban bus stations during operation. The consumption calculation module is used to calculate the predicted cycle consumption based on the unit cycle consumption, the vehicle operating intensity, and the preset intensity coefficient. The cycle calculation module is used to calculate the total inventory based on the existing inventory quantity and the quantity in transit, and to calculate the inventory support cycle based on the total inventory and the predicted cycle consumption. The status determination module is used to compare the inventory support period with the average delivery time, and determine whether the current site is in a state of insufficient support based on the comparison result. The generation module is used to calculate the replenishment lead time based on the insufficient support status of the site, the average delivery time and the inventory support cycle, and generate a purchase requisition instruction or a transfer instruction between sites based on the replenishment lead time, the total inventory and the predicted consumption period. The deviation calculation module is used to obtain the actual spare parts consumption of the operation process within a preset statistical period, and to calculate the consumption deviation value based on the actual spare parts consumption and the predicted period consumption. The correction module is used to correct the preset intensity coefficient based on the deviation comparison result between the consumption deviation value and the preset deviation range.

2. The IoT platform supply chain management system based on multi-source data analysis according to claim 1, characterized in that, The consumption calculation module includes: an intensity correction unit, which calculates an intensity correction consumption rate based on the unit cycle consumption and the vehicle operating intensity; a prediction calculation unit, which calculates a predicted cycle consumption based on a preset prediction cycle and the intensity correction consumption rate; and a prediction calculation unit, which calculates the predicted cycle consumption based on a historical fluctuation coefficient and the predicted cycle consumption, wherein the historical fluctuation coefficient is determined based on the average value and standard deviation of the predicted cycle consumption over a preset historical period.

3. The IoT platform supply chain management system based on multi-source data analysis according to claim 1, characterized in that, The cycle calculation module includes: an inventory aggregation unit, which calculates the sum of the existing inventory quantity and the quantity in transit to obtain the total inventory; a daily consumption conversion unit, which calculates the ratio of the predicted cycle consumption to the preset prediction cycle to obtain the daily predicted consumption; and a cycle calculation unit, which calculates the ratio of the total inventory to the daily predicted consumption to obtain the inventory support cycle.

4. The IoT platform supply chain management system based on multi-source data analysis according to claim 1, characterized in that, The status determination module includes: a comparison unit, which compares the inventory support period with the average delivery time to obtain the comparison result; and a status determination unit, which determines that the current status is one of insufficient site support when the comparison result is that the inventory support period is less than the average delivery time.

5. The IoT platform supply chain management system based on multi-source data analysis according to claim 1, characterized in that, The generation module includes a replenishment calculation unit, which is used to calculate the difference between the average delivery time and the inventory support period to obtain the replenishment lead time. An inventory status determination unit is used to determine the inventory status based on the total inventory and the predicted periodic consumption. The generation unit is used to generate the purchase request instruction or the transfer instruction based on the observation duration fluctuation and the inventory status, wherein the observation duration fluctuation is determined based on the replenishment advance time within a preset generation cycle.

6. The IoT platform supply chain management system based on multi-source data analysis according to claim 5, characterized in that, The inventory status determination unit includes: a redundancy calculation subunit, which is used to calculate the inventory redundancy rate based on the total inventory and the predicted periodic consumption; and a status determination subunit, which is used to determine that the inventory status is a pending allocation status when the inventory redundancy rate is greater than a preset redundancy threshold, and to determine that the inventory status is a low redundancy status when the inventory redundancy rate is less than or equal to the preset redundancy threshold.

7. The IoT platform supply chain management system based on multi-source data analysis according to claim 5, characterized in that, The generation unit includes: a duration statistics subunit, which is used to count all replenishment advance durations that are greater than a preset advance duration threshold within the preset generation cycle, and mark them as generation observation durations; a duration distribution calculation subunit, which is used to calculate the standard deviation of all the generation observation durations to obtain the observation duration volatility; and a generation subunit, which is used to generate the purchase request instruction or the transfer instruction based on the comparison results of the observation duration volatility and the respective thresholds of the inventory status.

8. The IoT platform supply chain management system based on multi-source data analysis according to claim 7, characterized in that, The generation subunit is used to determine that critical spare parts procurement is required when the observation duration fluctuation is less than or equal to a preset duration fluctuation threshold and the inventory status is the low redundancy status, so as to generate the procurement request instruction; and to determine that inter-site transfer is required when the observation duration fluctuation is greater than the preset duration fluctuation threshold and the inventory status is the pending allocation status, so as to generate the transfer instruction.

9. The IoT platform supply chain management system based on multi-source data analysis according to claim 1, characterized in that, The deviation calculation module includes: an actual acquisition unit, which is used to acquire the actual spare parts consumption during the operation within a preset statistical period; and a deviation rate calculation unit, which is used to calculate the absolute value of the relative deviation between the sum of all actual spare parts consumption and the consumption in the predicted period, so as to obtain the consumption deviation value.

10. The IoT platform supply chain management system based on multi-source data analysis according to claim 1, characterized in that, The correction module includes: an interval determination unit, which compares the consumption deviation value with the preset deviation interval to obtain the deviation comparison result; and a coefficient correction unit, which corrects the preset intensity coefficient based on the deviation comparison result and the consumption deviation value.

Images