A production line scheduling control method and system based on digital twinning

Abstract

This invention relates to the field of production line scheduling technology, and provides a production line scheduling and control method and system based on digital twins. The method includes constructing a digital twin model of the production line containing virtual equipment structures, task queues, accumulated running time, and performance degradation coefficients (health status parameters); real-time acquisition of equipment operating data to update the model's health parameters; prediction of equipment failure time by fitting parameter change curves through trend extrapolation; insertion of virtual maintenance events into the model to conduct scheduling simulations and calculate task completion times; and, in the event of task conflicts, combining equipment load and task urgency, redistributing conflicting tasks to suitable equipment using a load balancing strategy. This achieves dynamic perception of equipment health status and early prediction of failures, reserving reasonable maintenance space for scheduling, avoiding disruption of scheduling plans by sudden failures, solving load imbalance problems caused by temporary adjustments, and improving the overall operational stability and scheduling efficiency of the production line.

Description

Technical Field

[0001] This invention relates to the field of production line scheduling technology, and more specifically, to a production line scheduling and control method and system based on digital twins. Background Technology

[0002] The content in this section only provides background information related to this invention and may not constitute prior art.

[0003] In the process of modern manufacturing's transformation towards intelligent manufacturing, production line scheduling and control, as a core link to ensure efficient and orderly production, directly impacts production efficiency, product quality, and production costs. With the expansion of production scale and the increasing complexity of production processes, traditional production line scheduling methods are gradually being optimized by incorporating emerging technologies. Among these, digital twin technology, with its advantages of accurately mapping physical and virtual systems and enabling visualized monitoring of the production process, is widely used in production line management, providing a new technological path for scheduling and control. Through digital twin technology, the operating status of the production line can be simulated, monitored, and analyzed, thereby assisting in the formulation of scheduling decisions and attempting to solve problems such as insufficient coordination and delayed response in traditional scheduling methods.

[0004] However, existing digital twin-based production line scheduling and control technologies still have significant shortcomings. A prominent one is the failure to fully consider the dynamic changes in equipment health status and their potential impact on production scheduling during scheduling decisions and simulations. They also lack the ability to proactively predict and respond to equipment failures. Most existing technologies rely solely on real-time equipment operating data for scheduling adjustments, failing to anticipate declining trends in equipment health or predict potential failure times. This results in insufficient margin for equipment maintenance during scheduling plan development. When equipment malfunctions suddenly, the original scheduling plan is disrupted, leading to task allocation conflicts and production interruptions. Re-allocating tasks and adjusting the scheduling plan at this point often fails to achieve load balancing, impacting timely production delivery and potentially increasing production losses due to temporary adjustments. This reduces the overall operational stability and scheduling efficiency of the production line, failing to fully leverage the core advantages of digital twin technology in production line scheduling and control. Summary of the Invention

[0005] The purpose of this invention is to provide a production line scheduling and control method and system based on digital twins to improve the aforementioned problems. To achieve this objective, the technical solution adopted by this invention is as follows:

[0006] Firstly, this application provides a production line scheduling and control method based on digital twins, including:

[0007] Construct a digital twin model of the production line, which includes the structure of each virtual device on the production line, virtual task queues, and virtual parameters of device health status; the virtual parameters of device health status include cumulative running time and performance degradation coefficient.

[0008] Real-time equipment operation data is collected from the production line, including the continuous running time of the equipment, vibration data, and temperature data. Based on the continuous running time of the equipment, the cumulative running time of the corresponding virtual equipment is accumulated and updated. Based on the vibration data and temperature data, the performance degradation coefficient of the corresponding virtual equipment is calculated and updated.

[0009] The trend of the cumulative running time and the performance degradation coefficient is calculated, and the time point when the equipment health status value drops to the preset health threshold is predicted by the trend extrapolation algorithm, which is used as the fault prediction time. The trend extrapolation algorithm fits the corresponding change curves based on the historical cumulative running time data and historical performance degradation coefficient data, and calculates the time point when the health status value is lower than the preset health threshold by the trend extrapolation method.

[0010] Based on the updated digital twin model, a production scheduling simulation is performed. The simulation includes: inserting a virtual maintenance event of a preset duration into the running timeline of the corresponding virtual equipment model according to the fault prediction time; driving the digital twin model to simulate the task processing process in the virtual task queue according to the preset processing speed and current performance degradation coefficient of each virtual equipment model, and calculating the simulation completion time of each task.

[0011] When a virtual maintenance event causes a task conflict, the conflicting task is reassigned to another identical virtual device. The reassignment is based on the device's current load and the task's urgency, using a load balancing strategy to prioritize assigning the conflicting task to the identical virtual device with the lowest current load that can handle the task, and generating a corresponding scheduling scheme. The device's current load is obtained by calculating the total processing time of the tasks already assigned to the device. The task urgency is defined by the difference between the task's deadline and the current time.

[0012] Furthermore, the structure of the virtual device is constructed using 3D modeling software and imported into the digital twin model platform; each virtual device model is associated with a data interface to receive runtime data and drive the status update of its corresponding virtual device model.

[0013] Furthermore, the digital twin model platform also includes a visualization module, which displays the simulated task processing, virtual maintenance events, and equipment health status values ​​in real time in the form of animations and charts.

[0014] Furthermore, the step of calculating and updating the performance degradation coefficient of the corresponding virtual device based on vibration and temperature data specifically includes:

[0015] The amplitude and main frequency components are extracted from the vibration data. Based on the duration of the amplitude increase trend and the duration of the main frequency deviation from the reference range, the vibration characteristic value is determined. The duration of the temperature exceeding the reference temperature is extracted from the temperature data to determine the temperature characteristic value. Based on the proportion of the simultaneous increase of the vibration characteristic value and the temperature characteristic value, the decrease in the performance attenuation coefficient is increased. Based on the proportion of the simultaneous decrease of the vibration characteristic value and the temperature characteristic value, the decrease in the performance attenuation coefficient is decreased or kept unchanged.

[0016] Furthermore, the preset duration for virtual maintenance events is:

[0017] Based on the historical maintenance records of the corresponding equipment and the current value of the performance degradation coefficient, the preset duration is obtained by matching from the preset maintenance duration database.

[0018] Furthermore, the calculation of the current load of the device specifically includes:

[0019] For each virtual device, aggregate all assigned but uncompleted tasks in its virtual task queue, divide the standard processing time of each task by the device's current performance degradation factor to obtain the adjusted estimated processing time, and then add up all the adjusted estimated processing times to obtain the device's current load.

[0020] Furthermore, after generating the corresponding scheduling scheme, it also includes:

[0021] The scheduling scheme is evaluated using preset evaluation indicators. If the evaluation indicator is greater than the preset value, it is adopted; otherwise, an early warning is issued. The evaluation indicator is a weighted sum of the total completion time and the equipment utilization rate. The total completion time is obtained by summing the completion times of all tasks in the expected completion time series. The equipment utilization rate is obtained by calculating the ratio of the equipment busy time to the total simulation time.

[0022] Secondly, this application also provides a production line scheduling and control system based on digital twins, including:

[0023] The twin model construction module is used to build a digital twin model of the production line. This model includes the structure of each virtual device on the production line, the virtual task queue, and virtual parameters of the device health status. The virtual parameters of the device health status include the cumulative running time and the performance degradation coefficient.

[0024] The synchronization module is used to collect equipment operation data from the production line in real time. The operation data includes the continuous running time of the equipment, vibration data, and temperature data. Based on the continuous running time of the equipment, the module accumulates and updates the cumulative running time value of the corresponding virtual equipment. Based on the vibration data and temperature data, the module calculates and updates the performance degradation coefficient of the corresponding virtual equipment.

[0025] The health prediction module is used to calculate the fault prediction time, including: calculating the changing trend of the cumulative running time value and the performance degradation coefficient, and predicting the time point when the equipment health status value drops to the preset health threshold through the trend extrapolation algorithm, which is used as the fault prediction time; the trend extrapolation algorithm fits the corresponding change curves based on the historical cumulative running time value data and historical performance degradation coefficient data, and calculates the time point when the health status value is lower than the preset health threshold through the trend extrapolation method.

[0026] The scheduling simulation module is used to perform production scheduling simulation based on the updated digital twin model. The simulation includes: inserting a virtual maintenance event of a preset duration into the running timeline of the corresponding virtual equipment model according to the fault prediction time; driving the digital twin model to simulate the task processing process in the virtual task queue according to the preset processing speed and current performance degradation coefficient of each virtual equipment model, and calculating the simulation completion time of each task.

[0027] The scheduling decision module is used to reassign conflicting tasks to other identical virtual devices when a virtual maintenance event is inserted, causing a task conflict. The reassignment is based on the current load of the device and the urgency of the task, and adopts a load balancing strategy to prioritize assigning conflicting tasks to the identical virtual devices with the lowest current load that can handle the task, and generates a corresponding scheduling scheme. The current load of the device is obtained by calculating the total processing time of the tasks already assigned to the device. The urgency of the task is defined by the difference between the task deadline and the current time.

[0028] Thirdly, this application also provides an electronic device, including:

[0029] Memory, used to store computer programs;

[0030] A processor is used to implement the method steps as described in the first aspect when executing a computer program.

[0031] Fourthly, this application also provides a readable storage medium on which a computer program is stored, which, when executed by a processor, implements the method steps of the first aspect.

[0032] The beneficial effects of this invention are as follows:

[0033] This invention first constructs a complete digital twin model of the production line, incorporating the structure of each virtual device on the production line, virtual task queues, and relevant virtual parameters reflecting the health status of the equipment. Then, it collects various operational data from the production line in real time. Based on the collected continuous runtime, the cumulative runtime values ​​of the corresponding virtual devices in the digital twin model are accumulated and updated. Simultaneously, combined with vibration and temperature data, the performance degradation coefficient of the virtual devices is calculated and updated. Through this real-time data interaction and model parameter update, precise synchronization between the digital twin model and the actual production line equipment health status is achieved. Further analysis of the changing patterns of the cumulative equipment runtime values ​​and performance degradation coefficients allows for the early prediction of equipment health status using a trend extrapolation algorithm, yielding predicted equipment failure times. This effectively compensates for the shortcomings of existing technologies in lacking early failure prediction capabilities, enabling early detection of equipment health degradation trends and providing reliable data support for reserving maintenance space in scheduling schemes. Based on the updated digital twin model, the simulation operation of production scheduling continues. According to the predicted failure time, a virtual maintenance event of a preset duration is inserted into the runtime timeline of the corresponding virtual equipment model. The digital twin model is then driven, and the processing speed of each virtual equipment model is simulated, along with the current performance degradation coefficient, to simulate the processing of various tasks in the virtual task queue. The simulation completion time of each task is calculated. By simulating the impact of maintenance events on production tasks in advance, this method effectively avoids the problem in existing technologies where sudden equipment failures disrupt the original scheduling scheme due to the lack of reserved maintenance space. When the insertion of a virtual maintenance event causes a task conflict, the conflicting tasks are reassigned to other identical virtual equipment capable of handling the task. During the reassignment process, the current load of the equipment and the urgency of the task are comprehensively considered, and a load balancing strategy is adopted to prioritize assigning conflicting tasks to identical virtual equipment with the lowest current load that can handle the task. This resolves the task conflict problem, achieves balanced distribution of equipment load, avoids production interruptions, ensures timely delivery of production tasks, reduces production losses caused by temporary scheduling adjustments, and improves the overall operational stability and scheduling efficiency of the production line. Attached Figure Description

[0034] Figure 1 A flowchart of a production line scheduling and control method based on digital twins provided by the present invention;

[0035] Figure 2 A schematic diagram of a production line scheduling and control system based on digital twin provided by the present invention;

[0036] Figure 3 This is a schematic diagram of the structure of an electronic device provided by the present invention.

[0037] In the diagram: 201, Twin Model Construction Module; 202, Synchronization Module; 203, Health Prediction Module; 204, Scheduling Simulation Module; 205, Scheduling Decision Module. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0039] like Figure 1 As shown in the figure, an embodiment of the present invention proposes a production line scheduling and control method based on digital twins, comprising:

[0040] S101, Construct a digital twin model of the production line. This model includes the structure of each virtual device on the production line, the virtual task queue, and virtual parameters of the device health status. The virtual parameters of the device health status include the cumulative running time and the performance degradation coefficient.

[0041] Specifically, the production line digital twin model constructed by this invention must fully include three core elements: the structure of each virtual device on the production line, the virtual task queue, and virtual parameters of the device health status. The virtual parameters of the device health status are specifically defined as the cumulative running time value and the performance degradation coefficient. These two parameters are set based on the inherent correlation between device health status, running time, and performance loss. The cumulative running time value quantifies the total operating load of the device, while the performance degradation coefficient characterizes the degree of performance loss after long-term operation. Their synergistic effect accurately reflects the real-time health status of the device, providing crucial data support for subsequent fault prediction and scheduling optimization. For example, in a CNC lathe on an automotive parts production line, the corresponding virtual device structure must completely replicate the core components of the physical CNC lathe, such as the spindle, turret, and feed mechanism. The virtual task queue must synchronously record the parts processing tasks, task priorities, and processing requirements of the CNC lathe. The cumulative running time value in the virtual parameters of the device health status is initially set to 0, increasing by 1 for every hour of operation. The initial value of the performance degradation coefficient is... Set to 1.0 (no performance degradation state). As the equipment continues to run, this coefficient gradually decreases. When it drops to 0.7, it indicates that the equipment performance has significantly degraded and needs to be included in the maintenance plan.

[0042] Regarding the construction and import of virtual equipment structures, the specific implementation method involves using professional 3D modeling software (such as SolidWorks, UG, and other commonly used industrial 3D modeling tools) to construct a virtual equipment structure identical to the actual equipment. After construction, the 3D virtual model is imported into a pre-set digital twin model platform, realizing the digital implementation of the virtual equipment structure. Simultaneously, to achieve real-time synchronization between the virtual model and the equipment, each virtual equipment model needs to be associated with an independent data interface. The core function of this data interface is to act as a data transmission bridge between the equipment and the virtual model, receiving real-time operating data from the equipment (such as runtime, vibration data, and temperature data mentioned in subsequent steps), and converting this data into instructions to drive the virtual equipment model's state updates. This achieves synchronized linkage between the virtual equipment model's state and the equipment's operating state, ensuring that the virtual model can mirror the equipment's operating posture and working status in real time, preventing the virtual model from becoming disconnected from the equipment. For example, after the three-dimensional virtual model of the CNC lathe is constructed, it is imported into the TwinBuilder digital twin platform. An RS485 data interface is associated with the virtual CNC lathe model. This interface receives real-time operating data such as the spindle speed, turret displacement, and machining progress of the physical CNC lathe. When the spindle speed of the physical CNC lathe is adjusted from 1000 r / min to 1500 r / min, the data interface receives the data and immediately drives the spindle speed of the virtual CNC lathe model to be synchronously adjusted to 1500 r / min, so as to achieve real-time synchronization of status.

[0043] In addition, the digital twin model platform has a dedicated visualization module. This module is designed based on data visualization technology, which transforms abstract data such as the simulated task processing, virtual maintenance events, and equipment health status values ​​in the virtual model into intuitive and easy-to-understand animations and charts. This allows for real-time synchronous display, enabling staff to quickly and intuitively grasp the real-time operating status of the production line. Without needing to delve into complex underlying data, staff can clearly understand the task execution progress, maintenance needs, and health status of each piece of equipment, improving the efficiency and convenience of production line operation monitoring. At the same time, it provides an intuitive reference for scheduling decisions.

[0044] S102: Collect real-time operating data of the equipment from the production line. The operating data includes the continuous running time of the equipment, vibration data, and temperature data. Based on the continuous running time of the equipment, accumulate and update the cumulative running time value of the corresponding virtual equipment. Based on the vibration data and temperature data, calculate and update the performance degradation coefficient of the corresponding virtual equipment.

[0045] Specifically, the operating status of equipment is intuitively reflected through physical quantities such as operating time, vibration, and temperature. These data are directly and intrinsically related to the health status and performance degradation of the equipment, and are the core basis for quantifying the health status of the equipment. Among them, the continuous operating time of the equipment in this instance refers to the continuous running time of the equipment from the start of this operation to the time of data acquisition. If there is a shutdown or maintenance during the process, the timer is reset. This data is directly related to the cumulative operating load of the equipment and is the core basis for subsequent updates of the cumulative operating time value. Vibration data is the dynamic data generated by the friction and vibration of the components during the operation of the equipment. It can intuitively reflect the abnormal state of internal components such as wear and loosening. For example, bearing wear will lead to increased vibration amplitude and deviation of the main frequency, which is a key indicator for judging the performance degradation of the equipment. Temperature data refers to the real-time temperature of the machine body and core components (such as the spindle and motor) during the operation of the equipment. Excessive temperature usually means abnormal heat dissipation and accelerated component wear, which will accelerate the performance degradation of the equipment. Therefore, using these three data as the core collection objects can comprehensively and accurately capture the real-time operating status of the equipment and avoid misjudgment of the status caused by single data collection.

[0046] After data collection, the cumulative runtime of the corresponding virtual device is accumulated and updated based on the device's current continuous runtime. The principle is that the cumulative runtime is a core parameter for quantifying the total operating load of the device. The total operating load is positively correlated with the runtime; the longer the cumulative runtime, the more severe the wear and performance degradation of the device components. Therefore, by accumulating the current continuous runtime, the total operating load data of the device can be updated in real time, ensuring that the cumulative runtime of the virtual device is consistent with the actual operating load of the physical device, accurately reflecting the cumulative operating status of the device. The specific implementation process is as follows: After receiving the current continuous runtime data transmitted by the physical device, the digital twin model platform adds this data to the current cumulative runtime of the corresponding virtual device to obtain the updated cumulative runtime value. This updated value is then synchronously updated to the virtual health status parameters of the virtual device in the digital twin model, achieving dynamic iteration of the cumulative runtime value. The update of the cumulative runtime value uses discrete-time recursion, resulting in the formula:

[0047] (1)

[0048] In the formula, In the first At each data acquisition moment, the device Updated cumulative runtime; In the first The cumulative running time of each acquisition moment (i.e., the previous moment); In the first To the During the collection period, the equipment The duration of this continuous run.

[0049] Finally, based on vibration and temperature data, the performance degradation coefficient of the corresponding virtual device is calculated and updated. The principle is that the performance degradation of the device is closely related to the vibration and temperature states. Abnormal vibration (increased amplitude, shift in dominant frequency) and excessive temperature (exceeding the reference temperature) will accelerate the wear and aging of the device components, leading to a decline in device performance. Therefore, by extracting the characteristic parameters of vibration and temperature data and analyzing their changing trends, the speed and degree of device performance degradation can be accurately determined. This allows for dynamic adjustment of the update range of the performance degradation coefficient, enabling the performance degradation coefficient to accurately and in real-time characterize the performance loss state of the device and avoid performance characterization deviations caused by a fixed degradation rate.

[0050] The detailed implementation process must follow the logic of "feature extraction - feature analysis - coefficient adjustment" to ensure the scientific nature of the calculation process and the accuracy of the results. The first step is to extract vibration and temperature characteristic values: From the collected vibration data, two core characteristic parameters, amplitude and dominant frequency component, are extracted through data filtering and feature extraction algorithms (such as Fourier transform). Amplitude reflects the intensity of vibration, and the dominant frequency component reflects the frequency characteristics of vibration. The combination of these two parameters can comprehensively characterize the vibration state of the equipment. Subsequently, based on the duration of the amplitude increase trend and the duration of the dominant frequency deviation from the baseline range, the vibration characteristic value is determined through weighted calculation (the weighting coefficient can be preset according to the type of equipment on the production line; for example, the weight of the amplitude increase duration is set to 0.6, and the weight of the dominant frequency deviation duration is set to 0.4). The vibration characteristic value ranges from 0 to 1; a larger value indicates a more severe vibration anomaly and a faster performance degradation rate. Simultaneously, the duration of temperature exceeding the preset benchmark temperature is extracted from the collected temperature data. This duration is then normalized to determine temperature characteristic values ​​(the normalized temperature characteristic values ​​also range from 0 to 1). The longer the duration of temperature exceeding the benchmark temperature, the larger the temperature characteristic value, indicating more severe performance degradation due to excessive temperature. Subsequently, the rate of decrease in the performance degradation coefficient is adjusted based on the proportion of characteristic value change: the initial value of the performance degradation coefficient is set to 1.0 (indicating no performance degradation and optimal operating condition). As the equipment continues to operate, this coefficient gradually decreases, with the rate of decrease positively correlated with the rate of performance degradation; a larger decrease indicates faster performance degradation. When the coefficient drops to a preset threshold (e.g., 0.7), it indicates significant performance degradation, requiring inclusion in the maintenance plan. The specific adjustment rules are as follows: Based on the proportion of simultaneous increase in vibration characteristic value and temperature characteristic value, increase the decrease in performance attenuation coefficient. The principle is that simultaneous increase in both indicates a dual anomaly in the equipment (vibration anomaly + temperature anomaly), and the performance attenuation rate will be significantly accelerated. Therefore, it is necessary to increase the decrease to accurately match the actual performance loss. Based on the proportion of simultaneous decrease in vibration characteristic value and temperature characteristic value, decrease the decrease in performance attenuation coefficient or keep it unchanged. The principle is that simultaneous decrease in both indicates that the equipment's operating state is becoming more stable and the performance attenuation rate is slowing down. Therefore, it is necessary to reduce the decrease to avoid over-evaluating performance loss. If one characteristic value increases and the other decreases, the original decrease or a slight adjustment is maintained based on a comprehensive judgment of the changes in both.

[0051] When updating the performance degradation coefficient, the formula for calculating the vibration characteristic value is:

[0052] (2)

[0053] In the formula, For the first One cycle device The vibration characteristic value, ranging from [0,1]; In the first One collection cycle Within, the duration for which the vibration amplitude exceeds the reference value; In the first One collection cycle Within, the duration of the vibration's dominant frequency deviating from the reference range; A fixed data collection cycle duration; , These are the weights for amplitude and frequency, respectively.

[0054] The formula for calculating the temperature characteristic value is:

[0055] (3)

[0056] In the formula, For the first One cycle device The temperature characteristic value, ranging from [0,1]; In the first The duration during which the temperature exceeds the reference temperature within each acquisition cycle.

[0057] The formula for calculating the performance degradation coefficient is:

[0058] (4)

[0059] in,

[0060] (5)

[0061] In the formula, No. Performance degradation coefficient after each cycle update; This represents the performance degradation factor before the update. This is the attenuation adjustment amount for this cycle; The basic attenuation rate constant; , These are the weighting coefficients for the influence of vibration and temperature characteristic changes on attenuation, respectively.

[0062] S103, calculate the changing trends of the cumulative running time value and the performance degradation coefficient, and predict the time point when the equipment health status value drops to the preset health threshold through the trend extrapolation algorithm, which is used as the fault prediction time; the trend extrapolation algorithm fits the corresponding change curves based on the historical cumulative running time value data and the historical performance degradation coefficient data, and calculates the time point when the health status value is lower than the preset health threshold through the trend extrapolation method.

[0063] Specifically, since the cumulative running time and performance degradation coefficient work together to characterize the health status of equipment, accurate prediction of equipment failures requires first accurately capturing the changing trends of these two parameters and clarifying their changing patterns over time. For example, whether the parameters change at a constant rate, accelerate, or decelerate. This changing pattern directly determines the rate of decay of the equipment's health status, and thus the approximate time range for failure occurrence. Based on this, a trend extrapolation algorithm is used to predict the evolution of the equipment's health status, ultimately determining the time point when the equipment's health status value drops to a preset health threshold. This time point is then used as the failure prediction time. The preset health threshold is a critical value preset based on the actual operating requirements and maintenance standards of the production line equipment in this embodiment of the invention. When the equipment's health status value is below this threshold, it indicates that the equipment is close to a failure state. If maintenance is not carried out in time, a failure is highly likely to occur, affecting production progress and product quality. Therefore, accurate acquisition of the failure prediction time is a key prerequisite for subsequent proactive maintenance and optimized scheduling.

[0064] Furthermore, the trend extrapolation algorithm used in this embodiment of the invention is based on the principle of utilizing the continuity and regularity of the development of things. By analyzing the changing trends of historical data, it fits a mathematical curve that can accurately reflect the changing patterns of parameters. Then, based on this curve, it extends outward to predict the value or state of the parameters at a certain point in the future. Specifically, in this step, the algorithm first needs to collect and organize the historical cumulative running time data and historical performance degradation coefficient data of the corresponding virtual device. This historical data needs to cover the parameter values ​​under different operating stages and different load states of the device to ensure the comprehensiveness and representativeness of the data, thus providing a guarantee for the accuracy of curve fitting. Subsequently, based on the collected data... Based on the historical accumulated operating time data, appropriate fitting methods (such as linear fitting, exponential fitting, and other common trend fitting methods) are used to fit the corresponding curve of the accumulated operating time changing over time. At the same time, based on the historical performance degradation coefficient data, the corresponding curve of the performance degradation coefficient changing over time is also fitted using the same fitting method. Through these two curves, the growth pattern of the accumulated operating time and the decline pattern of the performance degradation coefficient can be clearly understood. Then, by combining the correlation between the equipment health status value and these two parameters, the specific time point when the equipment health status value falls below the preset health threshold is further calculated using the trend extrapolation method. This time point is the fault prediction time.

[0065] The specific calculation process is as follows:

[0066] Assume the curve of the cumulative trend of running time is as follows The curve showing the trend of performance degradation coefficient is as follows: Since equipment aging or health decline is usually not linear, but rather the rate of degradation accelerates with increasing usage time, a model is used to define the health status value by multiplying the performance degradation coefficient trend by a time decay term. The calculation formula is as follows:

[0067] (6)

[0068] In the formula, This represents the decay effect of runtime on health status.

[0069] Therefore, fault prediction time It is a solution that satisfies the following equation:

[0070] (7)

[0071] In the formula, The preset health status threshold (e.g., 0.7).

[0072] S104, based on the updated digital twin model, performs production scheduling simulation; the simulation includes: inserting a virtual maintenance event of a preset duration into the running timeline of the corresponding virtual equipment model according to the fault prediction time; driving the digital twin model to simulate the task processing process in the virtual task queue according to the preset processing speed and current performance degradation coefficient of each virtual equipment model, and calculating the simulation completion time of each task.

[0073] Specifically, the first core step in this process is the insertion of virtual maintenance events. This involves inserting a virtual maintenance event of a preset duration into the runtime timeline of the corresponding virtual equipment model based on the fault prediction time calculated in step S103. The principle is that the fault prediction time is the critical point in time when the equipment's health status value drops to a preset health threshold. If maintenance is not performed before this time, the equipment is highly likely to malfunction, leading to production interruption. Therefore, inserting virtual maintenance events in advance based on the fault prediction time is essentially a digital simulation of the proactive equipment maintenance process. By pre-setting maintenance steps in the virtual scenario, the impact of maintenance operations on task execution progress can be predicted in advance, providing a basis for subsequent task conflict judgment and scheduling adjustments. This enables the coordinated adaptation of maintenance plans and production tasks, avoiding conflicts between maintenance operations and production tasks, while ensuring that maintenance operations can be executed in time before equipment malfunctions, thus guaranteeing stable equipment operation. For example, taking a virtual equipment model corresponding to a CNC lathe in an automotive parts production line as an example, assuming that the trend extrapolation algorithm in step S103 predicts that the failure prediction time of the virtual CNC lathe is 1200 hours of operation, that is, its health status value will drop to a preset health threshold at this time. At this time, according to the requirements of this step, a virtual maintenance event of preset duration is inserted into the time node corresponding to 1200 hours on the running timeline of the virtual CNC lathe model. During the execution of the virtual maintenance event, the virtual CNC lathe model will suspend the execution of all pending tasks in the virtual task queue, simulating the maintenance shutdown state of the physical equipment, and ensuring that the simulation process is completely consistent with the equipment maintenance scenario in actual production.

[0074] The preset duration for virtual maintenance events is determined by matching the historical maintenance records of the corresponding equipment with the current value of the performance degradation coefficient from a preset maintenance duration database. The principle is that the maintenance duration of an equipment is directly related to the degree of performance degradation. The lower the performance degradation coefficient (the more severe the performance loss), the greater the maintenance workload and the longer the maintenance duration. Historical maintenance records provide practical data references for matching maintenance durations, ensuring that the preset duration closely matches the actual maintenance needs of the equipment. This avoids problems such as incomplete maintenance due to excessively short maintenance durations, leaving potential equipment malfunctions, or wasted production efficiency due to excessively long maintenance durations. The preset maintenance duration database is established in advance in this embodiment of the invention based on historical maintenance data, equipment model parameters, maintenance standards, and other information for all equipment on the production line. The database stores maintenance duration ranges corresponding to different performance degradation coefficient intervals, enabling rapid and accurate matching of maintenance durations.

[0075] The second core component is the simulation execution of the task processing process and the calculation of the simulation completion time. Specifically, it drives the updated digital twin model, and based on the preset processing speed and current performance degradation coefficient of each virtual device model, simulates the task processing process in the virtual task queue and calculates the simulation completion time of each task. The principle is that the preset processing speed of each virtual device model is a baseline parameter preset based on the rated processing capacity of the physical device, representing the maximum task processing efficiency of the device in a state without performance degradation (performance degradation coefficient is 1.0). The current performance degradation coefficient reflects the actual performance loss of the device after a period of operation. The combination of the two can accurately calculate the current actual task processing speed of the device, that is, actual processing speed = preset processing speed × current performance degradation coefficient. Based on this actual processing speed, the task processing process is simulated, which can realistically simulate the task processing efficiency of the device in the current healthy state, and thus accurately calculate the simulation time of each task from start to finish.

[0076] S105, when a virtual maintenance event is inserted and causes a task conflict, the conflicting task is reassigned to other identical virtual devices; the reassignment is based on the current load of the device and the urgency of the task, and a load balancing strategy is adopted to prioritize the allocation of the conflicting task to the identical virtual device with the lowest current load that can handle the task, and a corresponding scheduling scheme is generated; the current load of the device is obtained by calculating the total processing time of the tasks already assigned to the device; the urgency of the task is defined by the difference between the task deadline and the current time.

[0077] Specifically, when a virtual maintenance event causes a task conflict, i.e. during the execution of the virtual maintenance event, the corresponding virtual device needs to suspend all pending tasks, so that the tasks already assigned in its virtual task queue cannot be executed as originally planned, thus causing a conflict with the task deadline and the task arrangements of other devices. At this time, the conflicting tasks need to be reassigned to other identical virtual devices. Here, "identical virtual devices" refers to virtual devices with the same structure as the original virtual devices and with the same task processing capabilities, so as to be compatible with the processing requirements of conflicting tasks, ensure that the task processing standards are consistent with the original devices, and avoid product quality deviations due to device compatibility issues.

[0078] The redistribution of conflicting tasks must be based on the current load of the devices and the urgency of the tasks, and a load balancing strategy should be adopted. The core principle of this strategy is to make the load of each virtual device more balanced by reasonably allocating tasks, so as to avoid some devices being overloaded and causing task delays, and some devices being idle and causing resource waste. Conflicting tasks are given priority to be assigned to the same virtual device with the lowest current load that can handle the task. This can ensure that conflicting tasks are executed quickly and maximize the overall device utilization, thereby generating a corresponding scheduling scheme.

[0079] The current load of a device is obtained by calculating the total processing time of the tasks assigned to the device. The specific calculation logic is as follows: For each virtual device, all assigned but uncompleted tasks in its virtual task queue are aggregated. The standard processing time of each task is divided by the current performance degradation coefficient of the device to obtain the adjusted estimated processing time. Then, all the adjusted estimated processing times are added together to obtain the current load of the device. The principle of this calculation method is that the performance degradation coefficient will affect the actual processing speed of the device. By using the ratio of the standard processing time to the degradation coefficient, the actual processing time of the task under the current device state can be accurately reflected, ensuring the accuracy of the load calculation.

[0080] The formula for calculating the current load of the equipment is as follows:

[0081] (8)

[0082] In the formula, For equipment Current load (in time units); For devices already assigned The task set; For the task exist The standard processing time when =1; where

[0083] (9)

[0084] In the formula, For virtual devices The preset processing speed or rated processing speed.

[0085] In the conflict task reassignment decision function, for the conflict tasks to be reassigned... Select target device for:

[0086] (10)

[0087] In the formula, A collection of virtual devices with the same model and capabilities as the original devices; For the task The deadline; This is the current simulation time; The maximum load among the same devices; This is the latest deadline for conflict-related tasks; This is a trade-off factor between load balancing and task urgency.

[0088] After the scheduling plan is generated, its feasibility is evaluated using preset evaluation indicators. The evaluation indicators are the weighted sum of the total completion time and equipment utilization rate. The total completion time is obtained by summing the completion times of all tasks in the expected completion time series, reflecting the overall production efficiency. The equipment utilization rate is obtained by calculating the ratio of the equipment's busy time to the total simulated time, reflecting the efficiency of resource utilization. When the evaluation indicators are greater than the preset values, it indicates that the plan can balance efficiency and resource utilization and can be adopted. Otherwise, an early warning is issued to remind staff to optimize and adjust, thereby avoiding the use of ineffective scheduling plans, reducing production risks, and ensuring the stable and efficient operation of the production line.

[0089] Based on the above solution, the application embodiments of the present invention are as follows:

[0090] The application scenario was selected as a batch machining production line for automotive engine cylinder blocks. This production line is a continuous production line, which includes 12 production equipment arranged in the order of the process. The actual application effect of the method of the present invention was verified by quantitative experimental data of the whole process and comparative analysis under the same working conditions.

[0091] This embodiment focuses on a 4-cylinder engine block machining production line in an automotive parts manufacturing plant. The production line consists of 12 core machines connected in series according to the machining process sequence, including redundant machines of the same model that are compatible with the tasks (A1, A4, A5, and A8 in the table below), to meet the task redistribution requirements. The specific process and equipment layout is shown in the table below:

[0092] Table 1: Process and Equipment Layout Table

[0093]

[0094] Among them, the redundant equipment groups that can be used for compatible tasks are A1 and A5 (same model CNC lathes, which can be used for roughing / finishing tasks of cylinder mounting surfaces, with consistent standard processing time per piece) and A4 and A8 (same model CNC horizontal boring machines, which can be used for roughing / finishing tasks of cylinder bores, with consistent standard processing time per piece).

[0095] The production task for this experiment is to process 200 4-cylinder engine blocks in batches, with a total deadline of 72 hours. The production line will operate continuously for 24 hours, with a single machine operating at a rated capacity of 98%.

[0096] The core preset parameters are shown in the table below:

[0097] Table 2: Core Parameter Table

[0098]

[0099] Based on the parameters given in Tables 1 and 2, the following implementation method is adopted:

[0100] S101 uses SolidWorks to complete 1:1 3D structural modeling of 20 devices, imports them into the TwinBuilder digital twin model platform, and associates each virtual device model with an independent OPCUA data interface to receive real-time operating data of the physical devices and drive the virtual model status to update synchronously.

[0101] The model has a built-in virtual task queue that synchronizes the process requirements, standard processing time, and task deadlines for 200 cylinder block machining tasks. The initial task allocation corresponds one-to-one with the production line process.

[0102] Initial task allocation for redundant equipment groups: A1 is assigned 200 cylinder block mounting surface roughing tasks, A5 is assigned 200 cylinder block mounting surface finishing tasks; A4 is assigned 200 cylinder bore rough boring tasks, and A8 is assigned 200 cylinder bore finishing tasks.

[0103] Configure health status virtual parameters for each virtual device, with the initial value set to: the initial value of accumulated runtime. The initial value of the performance degradation coefficient is 0. The value is 1.

[0104] S102, after the production line starts, according to the data collection cycle. For a 10-minute cycle, the continuous running time, vibration data, and temperature data of all equipment are collected, and the health parameters are updated in real time based on formulas (1) to (5). Taking a production line running continuously for 24 hours (k=144 collection cycles) as an example, the collected data and parameter calculation results of redundant equipment (A1, A4, A5, A8) are shown in the table below:

[0105] Table 3: Measured Data and Calculation Results

[0106]

[0107] Among the incompatible devices, A2-A3, A6-A7, and A9-A12 all had performance degradation coefficients in the range of 0.93-0.95 after 24 hours of operation, with no obvious abnormal degradation. Their vibration and temperature characteristic values ​​were all below 0.12, and there were no out-of-tolerance warnings.

[0108] S103, based on historical data from 24 hours of operation, fits the cumulative operating time change curves for each device. Performance degradation coefficient variation curve Health status values ​​are calculated using formulas (6) to (7). It is predicted that it will drop to a preset health threshold. The time point is 0.7, which is the fault prediction time. Among them, the fault prediction times for incompatible devices, calculated through fitting, are all greater than 150 hours for A2-A3, A6-A7, and A9-A12, exceeding the total deadline of this task by 72 hours. Therefore, there is no fault risk within the task cycle, and no maintenance event needs to be inserted in advance. The fault predictions for redundant equipment groups are shown in the table below:

[0109] Table 4: Redundant Equipment Failure Prediction Table

[0110]

[0111] As shown in Table 4, only the fault prediction time of A1 and A4 falls within the 72-hour task cycle, posing a fault risk that may affect the delivery of the current task. Virtual maintenance events need to be inserted into the scheduling simulation. The fault risks of the other equipment exceed the current task cycle, so no intervention will be made in this scheduling to avoid ineffective maintenance affecting production efficiency.

[0112] S104: Based on the fault prediction time, combined with the historical maintenance records of the corresponding equipment and the current value of the performance degradation coefficient, a preset maintenance duration is obtained from the preset maintenance duration database. A virtual maintenance event is inserted before the fault occurs to avoid sudden faults in advance. The specific settings are shown in the table below:

[0113] Table 5: Virtual Maintenance Event Data Settings Table

[0114]

[0115] The digital twin model is driven by a virtual device model. Based on the preset processing speed and current performance degradation coefficient of each virtual device model, the simulation execution of tasks in the virtual task queue is performed, and the simulation completion time of each task is calculated. The rule is: actual processing speed of the device = preset processing speed × current performance degradation coefficient. The simulation results show that: after a 2-hour virtual maintenance event is inserted into device A1, the 10 roughing tasks originally scheduled to be executed in the maintenance window are halted. Due to the performance degradation caused by the maintenance, a total of 32 roughing tasks cannot be completed by the originally planned 60-hour deadline, resulting in a conflict with the task deadline. The standard processing time for each conflicting task is 12 minutes, and the total standard processing time is 32 × 12 = 384 minutes = 6.4 hours. After a 2-hour virtual maintenance event is inserted into device A4, the 8 roughing tasks originally scheduled to be executed in the maintenance window are halted. Due to the performance degradation, a total of 28 roughing tasks cannot be completed by the originally planned 55-hour deadline, resulting in a task conflict. The standard processing time for each conflicting task is 15 minutes, and the total standard processing time is 28 × 15 = 420 minutes = 7 hours. For A5, A8 and all incompatible devices, no maintenance events were inserted, and tasks could proceed as initially planned without execution conflicts.

[0116] S105, for the conflicting tasks of devices A1 and A4, based on the current load of the devices and the urgency of the tasks, a load balancing strategy is adopted to complete the task redistribution. That is, through formulas (8) and (9), the current load calculation results of the compatible device group corresponding to the conflicting tasks are shown in the table below (simulation time is 24h node):

[0117] Table 6: Current Load Calculation Results for Compatible Device Groups Corresponding to Conflicting Tasks

[0118]

[0119] The urgency of a task is defined by the difference between its deadline and the current time. In this conflict, the deadlines for all processes in the conflicting tasks are 65 hours (65 hours is determined by working backwards from the given total delivery deadline of 72 hours, taking into account the rigid machining time of all downstream processes in the cylinder block machining production line, the time loss due to equipment performance degradation, and the safety buffer for production line fluctuations). The current simulation time is 24 hours, with a difference of 41 hours, classifying these as high-urgency tasks that must be prioritized for timely completion to avoid impacting downstream processes and the overall delivery cycle.

[0120] Therefore, the conflict task redistribution details are as follows:

[0121] The 32 conflicting roughing tasks in A1: 12 tasks are prioritized and assigned to the idle window after A1 maintenance is completed, and the remaining 20 tasks are assigned to the compatible device A5 of the same model; after the assignment, the load of A5 is updated to 2461.5+(20×12) / 0.975≈2707.7min, the load balancing rate of the device group is improved by 21.3%, and there is no risk of overload.

[0122] The 28 conflicting coarse-firing tasks in A4: 10 tasks are prioritized and assigned to the idle window after A4 maintenance is completed, and the remaining 18 tasks are assigned to the compatible device A8 of the same model; after assignment, the load of A8 is updated to 3073.8+(18×15) / 0.976≈3350.2min, the load balancing rate of the device group is improved by 18.6%, and there is no risk of overload.

[0123] Evaluating scheduling schemes based on prior data:

[0124] The evaluation metric is: Comprehensive Evaluation Metric = Total Completion Time × 0.4 + Equipment Utilization Rate × 0.6; where, the total completion time is obtained by summing the completion times of all tasks in the expected completion time series. Equipment utilization rate is obtained by calculating the ratio of equipment busy time to total simulated time. The total completion time for the entire batch of tasks is 69.2 hours (less than the 72-hour task deadline, meeting the delivery requirements). The average equipment utilization rate of the entire production line is 91.8%. The comprehensive evaluation metric is 82.76, which is greater than the preset evaluation threshold of 75; therefore, it meets the requirements, and this scheduling scheme is adopted.

[0125] Comparative example:

[0126] To verify the technical advantages of the method of this invention, a comparative model with the same working conditions was set up, and the industry-standard traditional digital twin scheduling method was adopted. The core logic is as follows: real-time monitoring of production line status and routine task scheduling are achieved only through the digital twin model, without incorporating virtual parameters of equipment health status, without performing early fault prediction and virtual maintenance simulation, and only temporarily performing task reallocation and emergency scheduling adjustments after a sudden equipment failure; the remaining production line equipment, production tasks, basic operating parameters, and initial task allocation are completely consistent with the embodiments of this invention. Details are as follows:

[0127] Production line operation and failure occurrence: The production line was operating normally according to the initial task allocation, without any advance maintenance plan; after 51 hours of operation, equipment A4 suddenly failed due to performance degradation and stopped; after 57 hours of operation, equipment A1 suddenly failed due to performance degradation and stopped, both of which triggered unplanned shutdowns.

[0128] Emergency repair and temporary dispatch: After the failure, emergency shutdown and repair were carried out. The actual downtime of A4 failure was 5 hours and the actual downtime of A1 failure was 6 hours, which far exceeded the preset planned maintenance time of the method of this invention. The failure shutdown caused the remaining 62 rough boring tasks of A4 and the remaining 58 rough turning tasks of A1 to be completely stopped. During the emergency dispatch, all tasks were temporarily redistributed to A8 and A5 equipment.

[0129] Scheduling execution results: Due to the lack of advance simulation and load balancing planning, equipment A8 and A5 experienced severe overload. The peak load of A8 reached 4600 min, and the equipment utilization rate plummeted to 76.5%. The peak load of A5 reached 4100 min, resulting in severe task queuing delays. Downstream processes experienced multiple shutdowns due to material shortages.

[0130] Final operational metrics: Total completion time for the entire batch of tasks: 79.5 hours, exceeding the 72-hour task deadline by 7.5 hours, indicating a significant delay in task delivery. Average equipment utilization rate across the entire production line: 80.2%, 11.6 percentage points lower than the method of this invention. Equipment load balancing rate: 59.7%, 30.2 percentage points lower than the method of this invention. The overall parameter comparison is shown in the table below:

[0131] Table 7: Comparison of Experimental Results

[0132]

[0133] Based on the above experiments, it can be seen that, compared with the traditional production line scheduling method based on digital twins, the present invention achieves a fundamental logical innovation from passive emergency adjustment to proactive prediction and optimization. By constructing a full-element digital twin model that integrates virtual equipment structure, full-process task queues, and equipment health status parameters, and relying on the real-time collected operating data of production line equipment to complete the dynamic synchronous update of health status, it accurately predicts the risk of equipment failure within the production task cycle. Through the simulation and deduction of virtual maintenance events in advance, it quantifies the chain reaction of maintenance actions on the production plan. Combined with the real-time load of equipment and the urgency of tasks, it completes the intelligent and accurate redistribution of conflicting tasks with a load balancing strategy. This completely solves the inherent defects of traditional scheduling technology, which cannot perceive the trend of equipment health decline in advance and cannot reserve a reasonable maintenance window in the scheduling plan. It avoids a series of problems such as sudden equipment failure disrupting the established scheduling plan, task conflicts caused by temporary emergency adjustments, production line load imbalance, production process interruption, and delivery cycle delays, effectively ensuring the continuous and stable operation of the production line and the scientific nature of scheduling decisions. This verification was only applied to production line scenarios with relatively short process chains. However, in production line environments with longer process chains and richer redundant equipment configurations that can accommodate more tasks, the scheduling optimization capabilities and application value of this invention will be significantly improved. It can achieve resource optimization and production efficiency improvement across the entire production line to a greater extent, fully releasing the core advantages of digital twin technology in the field of intelligent manufacturing production scheduling.

[0134] like Figure 2 As shown, based on the same inventive concept, this embodiment provides a production line scheduling and control system based on digital twins, including:

[0135] The twin model construction module 201 is used to construct a digital twin model of the production line. This model includes the structure of each virtual device on the production line, the virtual task queue, and virtual parameters of the device health status. The virtual parameters of the device health status include the cumulative running time value and the performance degradation coefficient.

[0136] The synchronization module 202 is used to collect equipment operation data from the production line in real time. The operation data includes the continuous running time of the equipment, vibration data, and temperature data. Based on the continuous running time of the equipment, the module accumulates and updates the cumulative running time value of the corresponding virtual equipment. Based on the vibration data and temperature data, the module calculates and updates the performance degradation coefficient of the corresponding virtual equipment.

[0137] The health prediction module 203 is used to calculate the fault prediction time, including: calculating the changing trend of the cumulative running time value and the performance degradation coefficient, and predicting the time point when the equipment health status value drops to the preset health threshold through the trend extrapolation algorithm, which is used as the fault prediction time; the trend extrapolation algorithm fits the corresponding change curves based on the historical cumulative running time value data and historical performance degradation coefficient data, and calculates the time point when the health status value is lower than the preset health threshold through the trend extrapolation method.

[0138] The scheduling simulation module 204 is used to perform production scheduling simulation based on the updated digital twin model. The simulation includes: inserting a virtual maintenance event of a preset duration into the running timeline of the corresponding virtual equipment model according to the fault prediction time; driving the digital twin model to simulate the task processing process in the virtual task queue according to the preset processing speed and current performance degradation coefficient of each virtual equipment model, and calculating the simulation completion time of each task.

[0139] The scheduling decision module 205 is used to reallocate conflicting tasks to other identical virtual devices when a virtual maintenance event is inserted, causing a task conflict. The reallocation is based on the current load of the device and the urgency of the task, and adopts a load balancing strategy to prioritize the allocation of conflicting tasks to the identical virtual devices with the lowest current load that can handle the task, and generates a corresponding scheduling scheme. The current load of the device is obtained by calculating the total processing time of the tasks already assigned to the device. The urgency of the task is defined by the difference between the task deadline and the current time.

[0140] like Figure 3 As shown, based on the same inventive concept, this embodiment provides an electronic device, including:

[0141] Memory 302 is used to store computer programs;

[0142] Processor 301 is used to implement the method steps as described in the first aspect when executing a computer program.

[0143] Based on the same inventive concept, this embodiment provides a readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, it implements the method steps as described in the first aspect.

[0144] It should be noted that the specific methods by which each module performs operations in the system described in the above embodiments have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0145] The above are merely preferred embodiments of the present invention and are not intended to limit the present 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 production line scheduling and control method based on digital twins, characterized in that, include: Construct a digital twin model of the production line, which includes the structure of each virtual device on the production line, virtual task queues, and virtual parameters of device health status; The virtual parameters of the device health status include the cumulative running time value and the performance degradation coefficient; Real-time equipment operation data is collected from the production line, including the equipment's continuous running time, vibration data, and temperature data. Based on the equipment's continuous running time, the cumulative running time value of the corresponding virtual equipment is accumulated and updated. Based on the vibration and temperature data, calculate and update the performance degradation coefficient of the corresponding virtual device; The step of calculating and updating the performance degradation coefficient of the corresponding virtual device based on the vibration data and temperature data specifically includes: The amplitude and main frequency components are extracted from the vibration data. Based on the duration of the amplitude increase trend and the duration of the main frequency deviation from the reference range, vibration characteristic values ​​are determined. The duration of the temperature exceeding the reference temperature is extracted from the temperature data to determine temperature characteristic values. The decrease in the performance degradation coefficient is increased according to the ratio of the simultaneous increase of the vibration characteristic values ​​and the temperature characteristic values. The decrease in the performance degradation coefficient is decreased or kept unchanged according to the ratio of the simultaneous decrease of the vibration characteristic values ​​and the temperature characteristic values. The changing trends of the cumulative running time and the performance degradation coefficient are calculated, and the time point when the equipment health status value drops to a preset health threshold is predicted by the trend extrapolation algorithm, which is used as the fault prediction time. The trend extrapolation algorithm fits the corresponding change curves based on the historical cumulative running time data and historical performance degradation coefficient data, and calculates the time point when the health status value is lower than the preset health threshold by the trend extrapolation method. Based on the updated digital twin model, a production scheduling simulation is performed; the simulation includes: inserting a virtual maintenance event of a preset duration into the running timeline of the corresponding virtual device model according to the fault prediction time; driving the digital twin model to simulate the task processing process in the virtual task queue according to the preset processing speed and current performance degradation coefficient of each virtual device model, and calculating the simulation completion time of each task. When a virtual maintenance event causes a task conflict, the conflicting task is reassigned to another identical virtual device. The reassignment is based on the device's current load and the task's urgency, and a load balancing strategy is adopted to prioritize assigning the conflicting task to the identical virtual device with the lowest current load that can handle the task, and a corresponding scheduling scheme is generated. The device's current load is obtained by calculating the total processing time of the tasks already assigned to the device. The task urgency is defined by the difference between the task deadline and the current time.

2. The production line scheduling and control method based on digital twins according to claim 1, characterized in that, The structure of the virtual device is constructed using 3D modeling software and imported into the digital twin model platform; each virtual device model is associated with a data interface, which is used to receive the running data and drive the corresponding virtual device model state update.

3. The production line scheduling and control method based on digital twins according to claim 1, characterized in that, The digital twin model platform also includes a visualization module, which displays the simulated task processing, virtual maintenance events, and equipment health status values ​​in real time using animations and charts.

4. The production line scheduling and control method based on digital twins according to claim 1, characterized in that, The preset duration of the virtual maintenance event is: Based on the historical maintenance records of the corresponding equipment and the current value of the performance degradation coefficient, a preset duration is obtained by matching from a preset maintenance duration database.

5. The production line scheduling and control method based on digital twins according to claim 1, characterized in that, The calculation of the current load of the device specifically includes: For each virtual device, all assigned but uncompleted simulation tasks in its virtual task queue are aggregated. The standard processing time of each task is divided by the current performance degradation coefficient of the device to obtain the adjusted estimated processing time. Then, all adjusted estimated processing times are added together to obtain the current load of the device.

6. The production line scheduling and control method based on digital twins according to claim 1, characterized in that, After generating the corresponding scheduling scheme, the process also includes: The scheduling scheme is evaluated using preset evaluation indicators. If the evaluation indicators are greater than the preset values, the scheme is adopted; otherwise, an early warning is issued. The evaluation indicators are the weighted sum of the total completion time and the equipment utilization rate. The total completion time is obtained by summing the completion times of all tasks in the expected completion time series. The equipment utilization rate is obtained by calculating the ratio of the equipment busy time to the total simulation time.

7. A production line scheduling and control system based on digital twins, based on the production line scheduling and control method based on digital twins as described in claim 1, characterized in that, include: The twin model construction module is used to build a digital twin model of the production line. This model includes the structure of each virtual device on the production line, the virtual task queue, and virtual parameters of the device health status. The virtual parameters of the device health status include the cumulative running time value and the performance degradation coefficient. The synchronization module is used to collect equipment operation data from the production line in real time. The operation data includes the continuous running time of the equipment in this operation, vibration data, and temperature data. Based on the continuous running time of the equipment in this operation, the module accumulates and updates the cumulative running time value of the corresponding virtual equipment. Based on the vibration and temperature data, calculate and update the performance degradation coefficient of the corresponding virtual device; The health prediction module is used to calculate the fault prediction time, including: calculating the changing trends of the cumulative running time value and the performance degradation coefficient, and predicting the time point when the device health status value drops to a preset health threshold using a trend extrapolation algorithm, which is used as the fault prediction time; the trend extrapolation algorithm fits the corresponding change curves based on historical cumulative running time data and historical performance degradation coefficient data, and calculates the time point when the health status value is lower than the preset health threshold using the trend extrapolation method. The scheduling simulation module is used to perform production scheduling simulation based on the updated digital twin model. The simulation includes: inserting a virtual maintenance event of a preset duration into the running timeline of the corresponding virtual device model according to the fault prediction time; driving the digital twin model to simulate the task processing process in the virtual task queue according to the preset processing speed and current performance degradation coefficient of each virtual device model, and calculating the simulation completion time of each task. The scheduling decision module is used to reassign conflicting tasks to other identical virtual devices when a virtual maintenance event is inserted, causing a task conflict. The reassignment is based on the current load of the device and the urgency of the task, and adopts a load balancing strategy to prioritize assigning conflicting tasks to the identical virtual devices with the lowest current load that can handle the task, and generates a corresponding scheduling scheme. The current load of the device is obtained by calculating the total processing time of the tasks already assigned to the device. The urgency of the task is defined by the difference between the task deadline and the current time.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement a production line scheduling and control method based on digital twins as described in any one of claims 1 to 6 when executing the computer program.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed by a processor, implements a production line scheduling and control method based on digital twins as described in any one of claims 1 to 6.

Images