Digital employee building method and system adopting digital twinning

Abstract

The invention discloses a digital employee building method and system adopting digital twinning, and relates to the technical field of digital twinning, and the method comprises the steps: building a production line mirror image model in a virtual space based on the inherent parameter information and real-time parameter information of a target production line; according to the target product model, X simulation production schemes of the target product model and X scheme fitness corresponding to the X simulation production schemes are obtained, and N simulation production schemes with the maximum scheme fitness are selected as N alternative production schemes; obtaining product quality parameters of products produced under the N alternative production schemes, and screening the N alternative production schemes to obtain a plurality of qualified production schemes; and obtaining cost coefficients of a plurality of qualified production schemes of the current target product model, and selecting an optimal production scheme from the plurality of qualified production schemes as a production scheme of the production. The technical problem of inaccurate production line parameter adjustment in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of digital twin technology, specifically to a method and system for building digital employees using digital twins. Background Technology

[0002] With the continuous development of industrial production, the requirements for the intelligence and efficiency of production lines are becoming increasingly higher. However, when faced with complex and ever-changing production demands, traditional production line parameter adjustments often rely on manual methods and experience within a limited parameter range, which can easily lead to inaccurate adjustments, resulting in unstable product quality and increased production costs. Summary of the Invention

[0003] This application provides a method and system for building digital employees using digital twins, which solves the technical problem of inaccurate production line parameter adjustment in the prior art.

[0004] The technical solution to the above-mentioned technical problems in this application is as follows: In a first aspect, this application provides a method for building a digital employee using digital twins, characterized in that the method includes: Based on the inherent and real-time parameter information of the target production line, a dynamically updated production line mirror model of the target production line is built in virtual space using digital twin technology. The production line mirror model includes a pre-trained production cost prediction plugin. Based on the target product model, and using its defined parameter range and production line mirror model, X simulated production schemes for the target product model are obtained, along with X scheme fitness values ​​corresponding to the X simulated production schemes. Based on the scheme fitness values, N simulated production schemes with the highest scheme fitness values ​​are selected from the X simulated production schemes as N alternative production schemes, and production is carried out according to the N alternative production schemes respectively. The product quality parameters of the products produced under N alternative production plans are obtained through analysis. Based on the product quality parameters, the N alternative production plans are screened to obtain multiple qualified production plans and the quality coefficients of multiple qualified production plans, which are updated in real time. Obtain the cost coefficients of multiple qualified production plans for the current target product model. Based on the cost coefficients and quality coefficients of the qualified production plans, select the optimal production plan from the multiple qualified production plans as the production plan for this production.

[0005] Secondly, this application provides a digital employee setup system using digital twins, comprising: The mirror model building module is used to build a dynamically updated mirror model of the target production line in a virtual space based on the inherent and real-time parameter information of the target production line and using digital twin technology. The production line mirror model includes a pre-trained production cost prediction plugin. The process parameter generation module is used to obtain X simulated production schemes for the target product model based on its limited parameter range and production line mirror model, as well as X scheme fitness corresponding to the X simulated production schemes. Based on the scheme fitness, N simulated production schemes with the largest scheme fitness are selected from the X simulated production schemes as N alternative production schemes, and production is carried out according to the N alternative production schemes respectively. The production plan screening module is used to analyze and obtain the product quality parameters of the products produced under N alternative production plans. Based on the product quality parameters, the N alternative production plans are screened to obtain multiple qualified production plans and the quality coefficients of multiple qualified production plans, and the results are updated in real time. The cost coefficient acquisition module is used to acquire the cost coefficients of multiple qualified production plans for the current target product model. Based on the cost coefficients and quality coefficients of the qualified production plans, the optimal production plan is selected from the multiple qualified production plans as the production plan for this production.

[0006] This application provides one or more technical solutions, which have at least the following technical effects or advantages: This application provides a method and system for building digital employees using digital twins. Based on the inherent and real-time parameter information of the target production line, a mirror model of the production line is built in virtual space. First, based on the target product model, X simulated production plans and X corresponding plan fitness values ​​are obtained for each of the X simulated production plans. The N simulated production plans with the highest fitness values ​​are selected as N candidate production plans. Product quality parameters of the products produced under these N candidate production plans are then obtained. These N candidate production plans are further screened to obtain multiple qualified production plans. Finally, the cost coefficients of the multiple qualified production plans for the current target product model are obtained, and the optimal production plan is selected from these qualified plans as the production plan for this production run.

[0007] Through the above technical solutions, based on digital twin technology, a digital technician is built to simulate the production process of different products in real time, quickly select the optimal production plan, and realize intelligent management of the production line and scientific optimization and adjustment of parameters, thereby improving production efficiency and product quality. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating a method for building a digital employee using digital twins, provided in an embodiment of this application. Figure 2 This is a schematic diagram of a digital employee building system using digital twins, provided in an embodiment of this application.

[0010] The components represented by each number in the attached diagram are explained below: The module includes: 11 for building a mirror model, 12 for generating process parameters, 13 for screening production plans, and 14 for obtaining cost coefficients. Detailed Implementation

[0011] This application provides a method and system for building digital employees using digital twins, which addresses the technical problem of inaccurate production line parameter adjustment in the prior art.

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0014] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0015] Example 1, as Figure 1 As shown in the figure, this application provides a method for building a digital employee using digital twins, including: S10: Based on the inherent parameter information and real-time parameter information of the target production line, a dynamically updated production line mirror model of the target production line is built in virtual space using digital twin technology. The production line mirror model includes a pre-trained production cost prediction plugin. In this embodiment, a mirror model of the production line is first built based on the inherent information and real-time parameter information of the target production line. The inherent information remains largely unchanged after the production line is constructed. The real-time parameter information includes the current operating status of the production line, such as variable parameters like machine speed, temperature of heated parts, dwell time of the workpiece at each process, and feed rate. By integrating the inherent and real-time parameter information using digital twin technology, a mirror model of the production line that is highly similar to the target production line and can be dynamically updated is built in virtual space.

[0016] The pre-trained production cost prediction plugin is trained on production data and accurately predicts production costs based on the real-time operation of the production line and relevant parameters of the production task.

[0017] Specifically, step S10 in the method provided in the application embodiment includes: Based on the physical characteristics of the target production line equipment, obtain the inherent parameter information of the equipment; Based on IoT devices deployed on the production line, real-time parameter information of the target production line is collected. Based on the inherent and real-time parameter information of the target production line, a dynamically updated mirror model of the production line is constructed in virtual space using digital twin technology.

[0018] In this embodiment, when building a production line mirror model, based on the physical characteristics of the target production line equipment, the inherent parameter information of the target production line is first collected. This inherent parameter information is obtained from the production line's design documents, equipment manuals, and other materials. The inherent parameter information includes the production line's length, shape, dimensions, and the variable parameters of each piece of equipment within the production line. Based on IoT devices on the production line, various variable parameters and environmental parameters on the target production line are collected in real time. Real-time parameter information of the production line is obtained through sensors, monitoring equipment, and other means. Environmental parameters include temperature, humidity, and air pressure.

[0019] Based on the collected inherent and real-time parameter information of the target production line, a digital twin platform and related algorithms are used to process and integrate the information, constructing the basic architecture of a dynamically updated production line mirror model in virtual space. A pre-trained production cost prediction plugin is integrated into the production line mirror model, enabling it to interact and collaborate with other modules within the model.

[0020] Digital twin technology is a technology that links and dynamically maps physical entities with virtual digital models in real time. It uses digital means to construct a virtual mirror image that is highly consistent with the physical object (such as production line, equipment, city, etc.) in terms of form, behavior, and state, so as to realize two-way interaction and real-time synchronization between the physical world and the virtual space.

[0021] During production line operation, real-time parameter information is continuously updated, and the production line mirror model is also dynamically updated accordingly. The production line mirror model can reflect the actual operating status of the target production line in real time, providing accurate data support for subsequent production plan formulation and optimization.

[0022] In summary, compared to existing technologies, by monitoring various parameters in the production line mirror model in real time, potential problems in the production process can be identified in a timely manner, and corresponding measures can be taken in advance for adjustment and optimization, thereby improving production efficiency and product quality and reducing production costs.

[0023] S20: Based on the target product model, and its limited parameter range and production line mirror model, obtain X simulated production schemes for the target product model, and X scheme fitness corresponding to the X simulated production schemes. Based on the scheme fitness, select N simulated production schemes with the largest scheme fitness from the X simulated production schemes as N alternative production schemes, and carry out production according to the N alternative production schemes respectively. In this embodiment, based on the target product model, the maximum permissible parameter range under normal process conditions, and a production line mirror model, X optional process parameters and X scheme fitness values ​​corresponding to X simulated production schemes are randomly generated under the constraints of the production line model and the limited parameter range, for further optimization. The maximum permissible parameter range can be directly obtained from process documents and technical manuals.

[0024] Based on the fitness of the X schemes, through simulated production, N simulated production schemes with the highest fitness are selected from the X schemes, that is, N better schemes are selected from the X schemes as N alternative production schemes, and trial production is carried out according to the N alternative production schemes respectively.

[0025] Specifically, step S20 in the method provided in the application embodiment includes: Obtain the target product model for orders to be manufactured; Then, based on the target product model, retrieve the range of process parameters required for producing the target product, and use this range as the limiting parameter range; Based on the defined parameter range, the production line mirror model is invoked to perform virtual production scheduling for the target product model, thereby obtaining X simulated production plans for the target product model. Based on the production line mirror model and the pre-trained production cost prediction plugin, the simulated production time and simulated production energy consumption of the X simulated production schemes are simulated and calculated. Based on the simulated production time and energy consumption data of the X simulated production schemes, the fitness of the X simulated production schemes is obtained by weighting. Based on the fitness of the proposed schemes, N simulated production schemes with the highest fitness are selected from X simulated production schemes as N alternative production schemes; where X and N are both positive integers, and X = 5N; The target product model is produced using N alternative production plans.

[0026] In this embodiment, the process first locates the orders to be produced and obtains the target product model for those orders. Then, based on the target product model, it retrieves the range of process parameters required to produce the target product from the production database or product specification document. This range serves as the limiting parameter range for producing the product, ensuring that the production process meets product quality requirements.

[0027] Based on the product's defined parameter range, a pre-built production line mirror model is invoked to perform virtual production scheduling for the target product model. During the virtual scheduling process, the model will combine the actual conditions of the production line, such as equipment uptime and capacity limitations, to generate X simulated production plans for the target product model.

[0028] After generating simulated production plans, simulation calculations are performed on X simulated production plans based on the production line mirror model and a pre-trained production cost prediction plugin. The simulation calculations include simulated production time and simulated production energy consumption. Simulated production time refers to the time required to complete the production task according to each simulated production plan, while simulated production energy consumption refers to the amount of energy consumed during the production process, such as electricity and gas.

[0029] X simulated production plans are evaluated. Based on the simulated production time and energy consumption data, a weighted calculation is performed to obtain the fitness of each of the X simulated production plans. The fitness of the plan comprehensively considers the two important factors of production time and energy consumption. Through reasonable weight allocation, it can more comprehensively reflect the advantages and disadvantages of each simulated production plan.

[0030] Based on the fitness of the schemes, N simulated production schemes with the highest fitness are selected from X simulated production schemes. Here, X and N are both positive integers, and X = 5N. The production scheme mainly consists of the parameter settings of various equipment on the production line. After selecting N alternative production schemes, each of the N alternative production schemes is used to produce the target product model on the actual production line. During the actual production process, the production status of each alternative production scheme is monitored in real time, and various data are collected, such as product quality data, actual production time, and actual production energy consumption, to provide a basis for subsequent production scheme selection.

[0031] The methods for obtaining the production cost prediction plugin include: Collect historical production data of the target production line as a training set. The historical production data includes at least the environmental parameters and equipment parameters of the target production line during the historical production process, as well as the production time and energy consumption under the corresponding environmental parameters and equipment parameters. The production cost prediction plugin is built using a neural network and trained using a training set until convergence. By inputting the equipment and environmental parameters required for a specific production plan into the production cost prediction plugin, the plugin can predict and output the simulated production time and energy consumption corresponding to that specific production plan.

[0032] In this embodiment, the collected historical production data of the target production line is compiled into a training set. The historical production data includes environmental parameters and equipment parameters of the target production line during historical production processes, as well as the production time and energy consumption under the corresponding environmental and equipment parameters. The function of the production cost prediction plugin is to predict the production time and energy consumption of the target model product under the input environmental parameters and equipment parameter information in the production plan, and then select the corresponding training data.

[0033] A production cost prediction plugin is built using machine learning algorithms and based on neural networks. The plugin is trained on a training set until convergence. Iterative training based on historical production data from the target production line yields a converged production cost prediction plugin. For example, the iterative training of the production cost prediction plugin can be achieved through the following technical path: 1. Data preparation: Collect historical production data from the target production line as the training set. This historical data should include at least the environmental and equipment parameters of the target production line during historical production processes, as well as the corresponding production time and energy consumption under those parameters. 2. Model building: The input layer has the number of nodes equal to the dimension of the input features. For example, if there are 10 features (environmental parameters + equipment parameters), the input layer contains 10 nodes. Set 1-3 hidden layers, adjusting the number of nodes in each layer experimentally (e.g., 64, 32, etc.). The activation function is ReLU. The number of nodes in the output layer equals the number of predicted targets. For example, predicting only time requires 1 node, while predicting both time and energy requires 2 nodes. The output layer generally does not use an activation function and directly outputs continuous values. 3. Model Training: The preprocessed dataset is divided into training and validation sets in an 8:2 ratio. Equipment and environmental parameters from historical production plans in the training set are used as input features to output the corresponding simulated production time and energy consumption. The corresponding prediction samples in the training set are used as supervision labels. The Adam optimizer and mean squared error (MSE) loss function are used to construct the training framework. The batch size is set to 32 and the total number of training rounds is 50. An early stopping mechanism (patience=5) is introduced. When the validation set loss does not decrease for 5 consecutive rounds, the training process is automatically terminated, resulting in a trained production cost prediction plugin. This effectively avoids model overfitting while ensuring that the model reaches convergence.

[0034] By inputting the equipment and environmental parameters required for a specific production plan into the pre-built and trained production cost prediction plugin, the simulated production time and energy consumption corresponding to that specific production plan can be predicted.

[0035] Furthermore, the product quality parameters of the products produced under N alternative production plans are analyzed and obtained, including: Track the products produced according to the N alternative production plans, and obtain production waste, factory quality inspection results and after-sales feedback results as product quality parameters.

[0036] In this embodiment, quality tracking is performed on products produced using N alternative production plans to obtain information on production waste, factory inspection results, and after-sales feedback. Production waste is a crucial indicator of product quality, reflecting the quantity and proportion of defective products during the production process. By statistically analyzing the quantity of scrap and defective products and the specific stages that generate waste during production, potential problems in the production process can be identified, such as equipment malfunctions, unreasonable processes, and poor raw material quality.

[0037] Factory quality inspection results are a direct test of product quality. Based on product quality standards and inspection specifications, various performance indicators and appearance quality of the product are tested. The inspection results clearly indicate whether the product meets quality requirements, whether there are any quality defects, and the type and extent of those defects. After-sales feedback results reflect the actual usage and quality status of the product from the user's perspective. Users may encounter various problems during product use, such as product malfunctions, unstable performance, and inconvenience.

[0038] Therefore, the obtained production waste, factory quality inspection results, and after-sales feedback results are used as product quality parameters to fully reflect the problems existing in the production process and facilitate subsequent optimization.

[0039] Furthermore, the products manufactured according to the N alternative production plans are tracked to obtain production waste, factory quality inspection results, and after-sales feedback results as product quality parameters, including: Obtain the proportion of waste products generated during the production process to the total production under the corresponding production plan, as the production waste situation; Under the corresponding production plan, obtain the proportion of unqualified products to the total production during the factory quality inspection process, and use it as the factory quality inspection result. Under the corresponding production plan, the proportion of the number of times after-sales quality feedback was received within a preset time period to the total number of productions is used as the after-sales feedback result; The values ​​of production waste, factory quality inspection results, and after-sales feedback results are weighted and summed, and the reciprocal is taken as the product quality parameter.

[0040] There will be normal waste and loss during the product manufacturing process, and defective products will also appear during inspection. Quality problems may also exist after the product is sold. All of these problems are related to the production plan.

[0041] In this embodiment, different weights are assigned to production waste, factory inspection results, and after-sales feedback results. For example, production waste directly reflects losses during the production process and has a significant impact on costs, so it can be assigned a relatively high weight; factory inspection results are direct inspections before the product enters the market; and after-sales feedback results reflect product quality from the perspective of actual user use.

[0042] Based on actual production conditions and experience, the weights of each indicator are reasonably determined. Assume the weight of production scrap / loss is 0.5, the weight of factory quality inspection results is 0.3, and the weight of after-sales feedback results is 0.2. Multiply the values ​​of production scrap / loss, factory quality inspection results, and after-sales feedback results by their respective weights, and sum them to obtain the comprehensive value.

[0043] The product quality parameter is obtained by taking the reciprocal of the comprehensive value. The product quality parameter is positively correlated with product quality; that is, the better the product quality, the higher the product quality parameter.

[0044] By calculating product quality parameters, the product quality of N alternative production plans is quantitatively evaluated. Based on the magnitude of the product quality parameters, the alternative production plans are ranked, and the plan with the higher product quality parameters is selected first.

[0045] S30: Analyze and obtain the product quality parameters of the products produced under N alternative production schemes. Based on the product quality parameters, screen the N alternative production schemes to obtain multiple qualified production schemes and the quality coefficients of multiple qualified production schemes, and update them in real time. In this embodiment, after obtaining the product quality parameters of the products produced under N alternative production plans, the N alternative production plans are screened based on the product quality parameters. The selected production plans that meet the conditions are considered as multiple qualified production plans.

[0046] The obtained qualified production plans and their quality coefficients are updated in real time.

[0047] Specifically, step S30 in the method provided in the application embodiment includes: Set the product quality parameter thresholds for the target product model; Eliminate production plans that do not meet the product quality parameter thresholds to obtain multiple qualified production plans for the target product model; Based on the corresponding quality parameters of multiple qualified production plans for the target product model, the quality coefficients of multiple qualified production plans for the target product model are calculated. We continuously analyze the after-sales feedback results and update the qualified production plans and their quality coefficients in real time.

[0048] In this embodiment, a quality parameter threshold is set, and production plans with product quality parameters greater than the threshold are selected. The quality coefficient is an important indicator for measuring the quality of qualified production plans. Production plans with product quality parameters lower than the target product model's threshold are eliminated, and the remaining production plans are considered as multiple qualified production plans for the target product model.

[0049] For a qualified production plan, the quality coefficient is determined based on the specific values ​​of its product quality parameters. For example, the product quality parameters are normalized, and the normalized values ​​are used as the quality coefficient. The higher the quality coefficient, the better the quality of the product produced by the qualified production plan; conversely, the lower the quality coefficient, the worse the product quality.

[0050] During the production process, the quality coefficient of the qualified production plan is updated in real time. As production continues, new data such as production waste, factory quality inspection results, and after-sales feedback are collected to recalculate product quality parameters and update the quality coefficient.

[0051] In summary, real-time updates to the quality coefficient can promptly reflect the actual quality status of a production plan, enabling more rational decisions in subsequent production. Production plans should be dynamically adjusted based on changes in market demand and product quality. If the quality coefficient of a qualified production plan continues to decline during the update process, it indicates a potential problem in the production process, requiring adjustments or optimizations, such as checking equipment operating status and adjusting process parameters. Simultaneously, for qualified production plans with high and stable quality coefficients, expanding their production scale can be considered to improve overall product quality and market competitiveness.

[0052] S40: Obtain the cost coefficients of multiple qualified production plans for the current target product model. Based on the cost coefficients and quality coefficients of the qualified production plans, select the optimal production plan from the multiple qualified production plans as the production plan for this production.

[0053] In this embodiment of the application, when obtaining the cost coefficients of multiple qualified production schemes for the current target product model, it is necessary to comprehensively consider multiple cost factors, including raw material costs, equipment usage costs, labor costs, and energy consumption costs. For each qualified production scheme, the specific values ​​of each cost in the production process are statistically analyzed in detail, and then the cost coefficient of the scheme is obtained according to a certain calculation method.

[0054] The cost coefficient reflects the overall cost performance of each qualified production plan. The lower the cost coefficient, the more advantageous the production plan is in cost control, and the more cost-effective it is in producing the target product; conversely, the higher the cost coefficient, the higher the cost of the production plan.

[0055] After obtaining the cost coefficients of multiple qualified production plans, the optimal production plan is selected from these plans by combining them with the previously obtained quality coefficients. The quality coefficient reflects the quality of the products produced by the production plan, while the cost coefficient reflects the cost control capability of the production plan.

[0056] Specifically, step S40 in the method provided in the application embodiment includes: Based on IoT devices deployed on the production line, the historical actual energy consumption and historical actual time consumption of multiple qualified production schemes for the current target product model are obtained; Based on the historical actual energy consumption and historical actual time consumption of multiple qualified production schemes for the target product model, multiple cost coefficients for multiple qualified production schemes are calculated. Obtain the order information for this production run, and set the cost coefficient weight and quality coefficient weight for this production run according to the order requirements; Based on the cost coefficient weight and quality coefficient weight, the cost coefficient and quality coefficient of multiple qualified production schemes are weighted and summed to obtain the priority of multiple schemes that correspond one-to-one with the multiple qualified production schemes. The qualified production plan with the highest priority is selected as the production plan for this production.

[0057] In this embodiment, IoT devices deployed on the production line collect actual energy consumption and actual time consumption data of multiple qualified production plans for the current target product model during historical production processes. This actual energy consumption and actual time consumption data are crucial for calculating cost coefficients.

[0058] When calculating cost coefficients, the relationship between energy consumption, time consumption, and cost is comprehensively considered. Different production schemes have different energy consumption and time consumption, resulting in varying impacts on costs. For example, some production schemes may have lower energy consumption but longer consumption, while others may have shorter consumption but higher energy consumption. Therefore, it is necessary to convert historical actual energy consumption and historical actual time consumption into specific cost coefficients.

[0059] After obtaining the order information for this production run, the cost and quality coefficient weights should be set appropriately based on the specific requirements of the order. Different orders may have different emphases on product quality and cost. If the order has high requirements for product quality, then the quality coefficient weight should be increased; if the order is more sensitive to cost, then the cost coefficient weight should be increased.

[0060] Based on the pre-set cost and quality coefficient weights, the cost and quality coefficients of multiple qualified production plans are weighted and summed to obtain a priority level for each plan, corresponding one-to-one with the qualified production plans. This priority level comprehensively considers both cost and quality, providing a more complete reflection of the merits of each qualified production plan.

[0061] Finally, the qualified production plan with the highest priority was selected as the production plan for this production run. This selection takes into account both product quality and cost control, maximizing production efficiency while meeting order requirements.

[0062] Through the above steps, compared with existing technologies, by using the inherent and real-time parameter information of the target production line, a production cost prediction plugin is built based on neural networks to reasonably predict the simulated production time and energy consumption corresponding to a specific production plan, screen qualified production plans, and then select the optimal production plan as the production plan for this production by combining cost coefficients and quality coefficients, thereby effectively improving production efficiency.

[0063] In summary, the embodiments of this application have at least the following technical effects: This application provides a method and system for building digital employees using digital twins. Based on the inherent and real-time parameter information of the target production line, a mirror model of the production line is built in virtual space. First, based on the target product model, X simulated production plans and X corresponding plan fitness values ​​are obtained. The N simulated production plans with the highest fitness values ​​are selected as N candidate production plans. Product quality parameters of the products produced under these N candidate plans are then obtained. These N candidate plans are further screened to obtain multiple qualified production plans. Finally, the cost coefficients of the multiple qualified production plans for the current target product model are obtained, and the optimal production plan is selected from these qualified plans as the production plan for this production run. Through the above technical solution, based on digital twin technology, a digital technician is built to simulate the production process of different products in real time, quickly selecting the optimal production plan. This achieves intelligent management of the production line and scientific optimization and adjustment of parameters, improving production efficiency and product quality.

[0064] Example 2, as Figure 2 As shown, based on the same inventive concept as the digital employee building method using digital twins provided in Embodiment 1, this application also provides a digital employee building system using digital twins, including: The mirror model building module 11 is used to build a dynamically updated mirror model of the target production line in a virtual space based on the inherent parameter information and real-time parameter information of the target production line and using digital twin technology. The mirror model of the production line includes a pre-trained production cost prediction plugin. The process parameter generation module 12 is used to obtain X simulated production schemes for the target product model based on its limited parameter range and production line mirror model, as well as X scheme fitness corresponding to the X simulated production schemes. Based on the scheme fitness, N simulated production schemes with the largest scheme fitness are selected from the X simulated production schemes as N alternative production schemes, and production is carried out according to the N alternative production schemes respectively. The production plan screening module 13 is used to analyze and obtain the product quality parameters of the products produced under N alternative production plans, and based on the product quality parameters, screen the N alternative production plans to obtain multiple qualified production plans and the quality coefficients of multiple qualified production plans, and update them in real time. The cost coefficient acquisition module 14 is used to acquire the cost coefficients of multiple qualified production schemes for the current target product model, and select the optimal production scheme from the multiple qualified production schemes based on the cost coefficients and quality coefficients of the qualified production schemes as the production scheme for this production.

[0065] In one embodiment, the mirror model building module 11 is specifically used for: Based on the physical characteristics of the target production line equipment, obtain the inherent parameter information of the equipment; Based on IoT devices deployed on the production line, real-time parameter information of the target production line is collected. Based on the inherent and real-time parameter information of the target production line, a dynamically updated mirror model of the production line is constructed in virtual space using digital twin technology.

[0066] In one embodiment, the process parameter generation module 12 is specifically used for: Obtain the target product model for orders to be manufactured; Based on the target product model, retrieve the range of process parameters required for producing the target product, and use this range as the limiting parameter range; Based on the defined parameter range, the production line mirror model is invoked to perform virtual production scheduling for the target product model, thereby obtaining X simulated production plans for the target product model. Based on the production line mirror model and the pre-trained production cost prediction plugin, the simulated production time and simulated production energy consumption of the X simulated production schemes are simulated and calculated. Based on the simulated production time and energy consumption data of the X simulated production schemes, the fitness of the X simulated production schemes is obtained by weighting. Based on the fitness of the proposed schemes, N simulated production schemes with the highest fitness are selected from X simulated production schemes as N alternative production schemes; where X and N are both positive integers, and X = 5N; The target product model is produced using N alternative production plans.

[0067] Furthermore, in one embodiment of the application, the method for obtaining the production cost prediction plugin includes: Collect historical production data of the target production line as a training set. The historical production data includes at least the environmental parameters and equipment parameters of the target production line during the historical production process, as well as the production time and energy consumption under the corresponding environmental parameters and equipment parameters. The production cost prediction plugin is built using a neural network and trained using a training set until convergence. By inputting the equipment and environmental parameters required for a specific production plan into the production cost prediction plugin, the plugin can predict and output the simulated production time and energy consumption corresponding to that specific production plan.

[0068] Furthermore, in one embodiment of the application, product quality parameters of the products produced under N alternative production schemes are analyzed and obtained, including: Track the products produced according to the N alternative production plans, and obtain production waste, factory quality inspection results and after-sales feedback results as product quality parameters.

[0069] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0070] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0071] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for building a digital employee using digital twins, characterized in that, The method includes: Based on the inherent and real-time parameter information of the target production line, a dynamically updated production line mirror model of the target production line is built in virtual space using digital twin technology. The production line mirror model includes a pre-trained production cost prediction plugin. Based on the target product model, and using its defined parameter range and production line mirror model, X simulated production schemes for the target product model are obtained, along with X scheme fitness values ​​corresponding to the X simulated production schemes. Based on the scheme fitness values, N simulated production schemes with the highest scheme fitness values ​​are selected from the X simulated production schemes as N alternative production schemes, and production is carried out according to the N alternative production schemes respectively. The product quality parameters of the products produced under N alternative production plans are obtained through analysis. Based on the product quality parameters, the N alternative production plans are screened to obtain multiple qualified production plans and the quality coefficients of multiple qualified production plans, which are updated in real time. Obtain the cost coefficients of multiple qualified production plans for the current target product model. Based on the cost coefficients and quality coefficients of the qualified production plans, select the optimal production plan from the multiple qualified production plans as the production plan for this production.

2. The method for building a digital employee using digital twins according to claim 1, characterized in that, Based on the inherent and real-time parameter information of the target production line, digital twin technology is used to build a dynamically updated mirror model of the target production line in virtual space, including: Based on the physical characteristics of the target production line equipment, obtain the inherent parameter information of the equipment; Based on IoT devices deployed on the production line, real-time parameter information of the target production line is collected. Based on the inherent and real-time parameter information of the target production line, a dynamically updated mirror model of the production line is constructed in virtual space using digital twin technology.

3. The method for building a digital employee using digital twins according to claim 1, characterized in that, Based on the target product model, and using its defined parameter range and production line mirror model, X simulated production plans for the target product model are obtained, along with X fitness values ​​corresponding to these X simulated production plans. Based on these fitness values, N simulated production plans with the highest fitness values ​​are selected from the X simulated production plans as N alternative production plans. Production is then carried out according to each of these N alternative production plans, including: Obtain the target product model for orders to be manufactured; Then, based on the target product model, retrieve the range of process parameters required for producing the target product, and use this range as the limiting parameter range; Based on the defined parameter range, the production line mirror model is invoked to perform virtual production scheduling for the target product model, thereby obtaining X simulated production plans for the target product model. Based on the production line mirror model and the pre-trained production cost prediction plugin, the simulated production time and simulated production energy consumption of the X simulated production schemes are simulated and calculated. Based on the simulated production time and energy consumption data of the X simulated production schemes, the fitness of the X simulated production schemes is obtained by weighting. Based on the fitness of the proposed schemes, N simulated production schemes with the highest fitness are selected from X simulated production schemes as N alternative production schemes; where X and N are both positive integers, and X = 5N; The target product model is produced using N alternative production plans.

4. The method for building a digital employee using digital twins according to claim 1, characterized in that, Methods for obtaining the production cost prediction plugin include: Collect historical production data of the target production line as a training set. The historical production data includes at least the environmental parameters and equipment parameters of the target production line during the historical production process, as well as the production time and energy consumption under the corresponding environmental parameters and equipment parameters. The production cost prediction plugin is built using a neural network and trained using a training set until convergence. By inputting the equipment and environmental parameters required for a specific production plan into the production cost prediction plugin, the plugin can predict and output the simulated production time and energy consumption corresponding to that specific production plan.

5. The method for building a digital employee using digital twins according to claim 1, characterized in that, The analysis yields product quality parameters for products produced under N alternative production plans, including: Track the products produced according to the N alternative production plans, and obtain production waste, factory quality inspection results and after-sales feedback results as product quality parameters.

6. A method for building a digital employee using digital twins according to claim 5, characterized in that, Track the products manufactured according to the N alternative production plans, and obtain production waste, factory quality inspection results, and after-sales feedback results as product quality parameters, including: Obtain the proportion of waste products generated during the production process to the total production under the corresponding production plan, as the production waste situation; Under the corresponding production plan, obtain the proportion of unqualified products to the total production during the factory quality inspection process, and use it as the factory quality inspection result. Under the corresponding production plan, the proportion of the number of times after-sales quality feedback was received within a preset time period to the total number of productions is used as the after-sales feedback result; The values ​​of production waste, factory quality inspection results, and after-sales feedback results are weighted and summed, and the reciprocal is taken as the product quality parameter.

7. A method for building a digital employee using digital twins according to claim 6, characterized in that, Based on product quality parameters, the N alternative production plans are screened to obtain multiple qualified production plans and their quality coefficients, which are updated in real time, including: Set the product quality parameter thresholds for the target product model; Eliminate production plans that do not meet the product quality parameter thresholds to obtain multiple qualified production plans for the target product model; Based on the corresponding quality parameters of multiple qualified production plans for the target product model, the quality coefficients of multiple qualified production plans for the target product model are calculated. We continuously analyze the after-sales feedback results and update the qualified production plans and their quality coefficients in real time.

8. A method for building a digital employee using digital twins according to claim 1, characterized in that, Obtain the cost coefficients of multiple qualified production plans for the current target product model. Based on the cost coefficients and quality coefficients of the qualified production plans, select the optimal production plan from the multiple qualified production plans as the production plan for this production, including: Based on IoT devices deployed on the production line, the historical actual energy consumption and historical actual time consumption of multiple qualified production schemes for the current target product model are obtained; Based on the historical actual energy consumption and historical actual time consumption of multiple qualified production schemes for the target product model, multiple cost coefficients for multiple qualified production schemes are calculated. Obtain the order information for this production run, and set the cost coefficient weight and quality coefficient weight for this production run according to the order requirements; Based on the cost coefficient weight and quality coefficient weight, the cost coefficient and quality coefficient of multiple qualified production schemes are weighted and summed to obtain the priority of multiple schemes that correspond one-to-one with the multiple qualified production schemes. The qualified production plan with the highest priority is selected as the production plan for this production.

9. A digital employee setup system using digital twins, characterized in that, For performing the method according to any one of claims 1-8, comprising: The mirror model building module is used to build a dynamically updated mirror model of the target production line in a virtual space based on the inherent and real-time parameter information of the target production line and using digital twin technology. The production line mirror model includes a pre-trained production cost prediction plugin. The process parameter generation module is used to obtain X simulated production schemes for the target product model based on its limited parameter range and production line mirror model, as well as X scheme fitness corresponding to the X simulated production schemes. Based on the scheme fitness, N simulated production schemes with the largest scheme fitness are selected from the X simulated production schemes as N alternative production schemes, and production is carried out according to the N alternative production schemes respectively. The production plan screening module is used to analyze and obtain the product quality parameters of the products produced under N alternative production plans. Based on the product quality parameters, the N alternative production plans are screened to obtain multiple qualified production plans and the quality coefficients of multiple qualified production plans, and the results are updated in real time. The cost coefficient acquisition module is used to acquire the cost coefficients of multiple qualified production plans for the current target product model. Based on the cost coefficients and quality coefficients of the qualified production plans, the optimal production plan is selected from the multiple qualified production plans as the production plan for this production.

Images