Intelligent production scheduling optimization system and method based on digital twinning

Abstract

The invention relates to the technical field of intelligent production, in particular to an intelligent production scheduling optimization system and method based on digital twinning, and the method comprises the steps: collecting a plurality of basic data, preprocessing the basic data to obtain preprocessed data, then building a production management model, inputting the preprocessed data into the production management model, and carrying out the optimization of the production management model. The method comprises the following steps: establishing a production management model, enabling the production management model to continuously learn a relationship between production data and employee data, obtaining a trained production management model, finally collecting production data demands, and obtaining a production plan based on the production data demands. And multiple items of basic data are used as influence factors of the production management model, so that the production plan obtained by the production management model is more comprehensive, and the rationality and reliability of the production plan can be ensured.

Description

Technical Field

[0001] The present application relates to the field of intelligent production technologies, and particularly to an intelligent production scheduling optimization system and method based on digital twin. Background Art

[0002] Intelligent manufacturing should include intelligent manufacturing technologies and intelligent manufacturing systems. The intelligent manufacturing system can not only continuously enrich the knowledge base in practice, but also has a self-learning function, and also has the ability to collect and understand environmental information and its own information, and perform analysis, judgment, and plan its own behavior.

[0003] Digital twin technology is increasingly used in the production process of intelligent manufacturing. Digital twin is a dynamic digital mapping system that relies on physical models, sensors, and real-time data to construct entity objects in a virtual space, and can realize the simulation, analysis, and optimization of the entire life cycle.

[0004] Chinese Patent with Publication No. CN115202294A discloses a manufacturing workshop production scheduling method based on digital twin, which optimizes the manufacturing workshop according to the digital twin model by real-time monitoring the state of the manufacturing workshop; adjusts the digital twin model according to the state of the manufacturing workshop; generates a new scheduling plan according to the operation assignment result and constraint conditions; calculates the machine deviation degree between the original scheduling plan and the new scheduling plan, and executes the new scheduling plan with the smallest machine deviation degree. However, in the prior art, only the production efficiency of machines is considered, while the impact of employees' production efficiency on the production plan is ignored, resulting in an imperfect plan in the actual planning process. Summary of the Invention

[0005] Based on this, in view of the defect that traditional Internet of Things multi-node data is difficult to handle flood peak data, it is necessary to propose an intelligent production scheduling optimization system and method based on digital twin.

[0006] On the one hand, the present application provides an intelligent production scheduling optimization system based on digital twin, including:

[0007] A production module;

[0008] A data acquisition module that acquires the basic data of the production module through the data acquisition module;

[0009] A management module that is communicatively connected to the data acquisition module and realizes the production plan management of the production module through the management module.

[0010] Preferably, the basic data includes equipment data, production data, and employee data.

[0011] On the other hand, the present application provides an intelligent production scheduling optimization method based on digital twin, including:

[0012] Collect multiple pieces of basic data, preprocess the basic data to obtain preprocessed data;

[0013] Create a production management model;

[0014] Input the preprocessed data into the production management model, enabling the production management model to continuously learn the relationship between production data and employee data, and obtain a trained production management model;

[0015] Collect production data requirements and obtain a production plan based on the production data requirements.

[0016] Preferably, collecting multiple pieces of basic data and preprocessing the basic data to obtain preprocessed data includes:

[0017] Create a basic database;

[0018] Set collection parameters; the collection parameters include collection location, collection frequency, and collection period;

[0019] Based on the collection parameters, collect equipment data, production data, and employee data respectively, and put all the collected data into the basic database; the equipment data includes the production data includes temperature data and energy consumption data, and the employee data includes employee efficiency;

[0020] Randomly select a piece of basic data from the basic database;

[0021] Judge whether there is missing data in the basic data. If there is missing data, fill in the data;

[0022] Return to randomly select a piece of basic data from the basic database until all the basic data in the basic database have been selected, and obtain preprocessed data.

[0023] Preferably, inputting the preprocessed data into the production management model, enabling the production management model to continuously learn the relationship between production data and employee data, and obtain a trained production management model includes:

[0024] Divide all the preprocessed data into a training set and a test set according to a random ratio;

[0025] Input the training set into the production management model to train the production management model until a trained production management model is obtained; the trained production management model has the ability to automatically output a corresponding production plan according to the input production demand data;

[0026] Input the test set into the trained production management model to verify whether the trained production management model is trained.

[0027] Preferably, input the training set into the production management model to train the production management model until the trained production management model is obtained, including:

[0028] Randomly select a basic data from the training set based on the collection location;

[0029] Input the equipment data into the production management model and construct a production model through the production management model;

[0030] Establish the coupling relationship between production data and employee data, and use the coupling relationship of production data-employee data as a training sample;

[0031] Return to randomly select a basic data from the training set based on the collection location until all the basic data in the training set are selected, and obtain multiple training samples.

[0032] Preferably, input the training set into the production management model to train the production management model until the trained production management model is obtained, and further include:

[0033] Select a training sample and input the training sample into the production management model;

[0034] Calculate the productivity of each production model through Formula 1;

[0035]

[0036] Among them, C is the productivity of the production model, Q i is the production efficiency of the i-th production equipment model, t is the working time of the i-th production equipment model, and k is the total number of production equipment models;

[0037] Statistical production efficiency of each employee;

[0038] Construct a production evaluation function through Formula 2 based on the productivity of the production model and the production efficiency of the employee;

[0039]

[0040] Among them, R is the production evaluation function, T is the total production time, N is the productivity of the model, V is the production cost, α is the evaluation coefficient of the total production time, β is the evaluation coefficient of the productivity of the model, γ is the evaluation coefficient of the production cost, and α + β + γ = 1.

[0041] Preferably, collect production data requirements and obtain a production plan based on the production data requirements, including:

[0042] Collect production demand data and input the production demand data into the trained production management model;

[0043] Obtain the production plan output by the trained production management model;

[0044] Conduct production scheduling according to the production plan.

[0045] Preferably, the production demand data includes one or more of production cycle, production model productivity, and production cost.

[0046] Preferably, the production model includes a production equipment model, a production line model composed of multiple production equipment models, and a factory model composed of multiple production line models.

[0047] By collecting a number of basic data, preprocessing the basic data to obtain preprocessed data, then creating a production management model, and inputting the preprocessed data into the production management model, enabling the production management model to continuously learn the relationship between production data and employee data, obtaining the trained production management model, and finally collecting production data requirements and obtaining the production plan based on the production data requirements. This application constructs a production management model through digital twin technology and uses a number of basic data as influencing factors of the production management model, thereby ensuring that the production plan obtained by the production management model is more comprehensive and can ensure the rationality and reliability of the production plan. Description of the Drawings

[0048] Figure 1 It is a schematic structural diagram of an intelligent production scheduling optimization system based on digital twin provided by an embodiment of the present application.

[0049] Figure 2 It is a flow chart of an intelligent production scheduling optimization method based on digital twin provided by an embodiment of the present application.

[0050] Reference Signs: 100, production module; 200, data acquisition module; 201, equipment side;

[0051] 202, employee side; 300, management module. Detailed Embodiment

[0052] In order to make the purpose, technical solutions and advantages of the present application clearer, the following further describes the present application in detail with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0053] On the one hand, the present application provides an intelligent production scheduling optimization system based on digital twin.

[0054] As Figure 1As shown, in an embodiment of the present application, an intelligent production scheduling optimization system based on digital twin includes a production module 100, a data acquisition module 200, and a management module 300. The basic data of the production module 100 is collected through the data acquisition module 200. The management module 300 is communicatively connected to the data acquisition module 200, and the production plan management of the production module 100 is realized through the management module 300.

[0055] Specifically, the data acquisition module 200 includes a device end 201 and an employee end 202. Among them, the device end 201 can collect the device data and production data of production equipment;

[0056] It should be noted that this embodiment relates to an intelligent production scheduling optimization system based on digital twin. Through the production module 100, production activities are carried out. On the one hand, the data acquisition module 200 can collect the device data of the production module 100 itself and the production data when the production module 100 is engaged in production activities. On the other hand, the data acquisition module 200 can also collect the relevant data of employees. After the data acquisition module 200 completes the collection of basic data, it will transmit the basic data to the management module 300. The management module 300 combines the basic data to control and manage the production module 100, so as to ensure the efficient operation of the production module 100.

[0057] In an embodiment of the present application, the basic data includes device data, production data, and employee data.

[0058] It should be noted that the device data includes static data of the device such as the length, width, and height of the device, so as to construct a production model based on the device data. The production data includes dynamic data during device operation such as device operating temperature and device operating voltage, so as to construct an operation model of the device by combining the production data and the production model. The employee data includes the work efficiency, working hours, and employee salary of employees.

[0059] As Figure 2 shown, the present application also provides an intelligent production scheduling optimization method based on digital twin, including:

[0060] S100, collecting multiple pieces of basic data and preprocessing the basic data to obtain preprocessed data;

[0061] S200, creating a production management model;

[0062] S300, inputting the preprocessed data into the production management model, so that the production management model continuously learns the relationship between production data and employee data to obtain a trained production management model;

[0063] S400, collect the production data requirements, and obtain the production plan based on the production data requirements.

[0064] It should be noted that by collecting a number of basic data, preprocessing the basic data to obtain preprocessed data, then creating a production management model, and inputting the preprocessed data into the production management model, enabling the production management model to continuously learn the relationship between production data and employee data, obtaining the trained production management model, and finally collecting the production data requirements and obtaining the production plan based on the production data requirements. In this application, a production management model is constructed through digital twin technology, and a number of basic data are used as influencing factors of the production management model, so as to ensure that the production plan obtained by the production management model is more comprehensive, and the rationality and reliability of the production plan can be guaranteed.

[0065] In an embodiment of the present application, S100 includes:

[0066] S110, create a basic database;

[0067] S120, set the collection parameters; the collection parameters include the collection location, collection frequency, and collection period;

[0068] S130, respectively collect equipment data, production data, and employee data based on the collection parameters, and put all the collected data into the basic database;

[0069] S140, randomly select a basic data from the basic database;

[0070] S150, determine whether there is missing data in the basic data. If there is missing data, fill in the data;

[0071] Specifically, when filling in the data, mean filling can be used;

[0072] S160, return to step S140 until all the basic data in the basic database have been selected, and obtain the preprocessed data.

[0073] It should be noted that after collecting a number of basic data according to the collection parameters, it is necessary to preprocess each item of basic data separately. In addition to checking for missing basic data, duplicate data removal operations can also be performed on the basic data. By preprocessing the basic data, the integrity of the basic data can be guaranteed, and thus the reliability in subsequent training of the production management model using the basic data can be improved.

[0074] In an embodiment of the present application, S300 includes:

[0075] S310, divide all the preprocessed data into a training set and a test set according to a random ratio;

[0076] Specifically, the division ratio of the training set should be greater than that of the test set to ensure that the amount of data contained in the training set is sufficient;

[0077] S320, input the training set into the production management model to train the production management model until the trained production management model is obtained; the trained production management model has the ability to automatically output the corresponding production plan according to the input production demand data;

[0078] S330, input the test set into the trained production management model to verify whether the trained production management model is trained successfully.

[0079] It should be noted that the preprocessed data is used as the training data of the production control model, so that the production management model continuously learns the corresponding relationship between production data, employee data and production plans, and finally obtains the trained production management model.

[0080] When verifying whether the trained production management model is trained successfully, one or more of the response time, accuracy rate and satisfaction of the output result of the trained production management model can be used as the evaluation criteria.

[0081] In an embodiment of the present application, S320 includes:

[0082] S321, randomly select a basic data from the training set based on the collection location;

[0083] S322, input the equipment data into the production management model and construct a production model through the production management model;

[0084] Specifically, construct a production model through the Unity3D engine;

[0085] S323, establish the coupling relationship between production data and employee data, and use the coupling relationship of production data-employee data as a training sample;

[0086] Specifically, since there is a corresponding employee at each production position on the production line, and the proficiency of different employees is different, resulting in different work efficiencies, it is necessary to correspond the employee data of each employee with the production data he operates;

[0087] S324, return to step S321 until all the basic data in the training set are selected, and obtain multiple training samples.

[0088] It should be noted that in the actual production process, each employee has a corresponding position. Therefore, it is necessary to align the employee data with their positions in the production model and associate the employee data with the production data, so as to take the employee data as an influencing factor for adjusting the production plan, and further ensure that the employees can be taken into account when the production management model outputs the production plan after training.

[0089] In an embodiment of the present application, S320 further includes:

[0090] S325, select a training sample and input the training sample into the production management model;

[0091] S326, calculate the productivity of each production model through Formula 1;

[0092]

[0093] Where C is the productivity of the production model, Q i is the production efficiency of the i-th production equipment model, t is the working time of the i-th production equipment model, and k is the total number of production equipment models;

[0094] S327, count the production efficiency of each employee;

[0095] Specifically, since the physical states of different employees are different, the production efficiencies of different employees are also different. When counting the production efficiency of employees, it can be counted according to different cycles. For example, for each employee, the production efficiency of each day, each month, and each quarter can be counted respectively;

[0096] S328, construct a production evaluation function through Formula 2 based on the productivity of the production model and the production efficiency of employees;

[0097]

[0098] Where R is the production evaluation function, T is the total production time, N is the productivity of the model, V is the production cost, α is the evaluation coefficient of the total production time, β is the evaluation coefficient of the productivity of the model, γ is the evaluation coefficient of the production cost, and α + β + γ = 1;

[0099] Specifically, the productivity of the model is the sum of the work efficiencies of all employees, that is Where N is the productivity of the model, M a is the production efficiency of the a-th employee, and B is the total number of employees;

[0100] The production cost mainly includes the material cost and the employee cost. In production, the material cost is generally determined by the material cost of the products to be produced, while the employee cost is generally determined based on the number of employees and the salary of each employee, that is where V is the production cost, B is the number of employees, P a is the salary of the a-th employee, and S is the material cost.

[0101] Optionally, in addition to the various influencing factors included in Formula 3, other influencing factors can also be added to the production evaluation function to improve the comprehensiveness of the production evaluation function.

[0102] It should be noted that the productivity of the production model is calculated through Formula 1. This productivity is the maximum production value that the production model can generate within time t, and the production efficiency of an employee is also the maximum workload that an employee can complete at their corresponding position within time t.

[0103] After obtaining the productivity of the production model and the production efficiency of the employees, an evaluation function for the entire production model can be created by combining these two. Since this evaluation function includes various influencing factors such as the productivity of the production model, the production efficiency of the employees, and the utilization rate of the production model, the production plan can be arranged by combining different focuses.

[0104] For example, the production demand data is 30,000 pieces, and the delivery time is 5 days. If the productivity of the existing production model is 10,000 pieces per day, the production efficiency of employee A is 3,000 pieces per day, the production efficiency of employee B is 4,000 pieces per day, the production efficiency of employee C is 3,000 pieces per day, the production efficiency of employee D is 2,000 pieces per day, and the salaries of multiple employees are 600, 700, 600, and 400 respectively. Then in this case, if the production plan requires the lowest labor cost, the production plan obtained after inputting the above data and requirements into the trained production control model is: employee B and employee D work for 5 days. And if the production plan requires the shortest production time, the production plan is: employee A, employee B, and employee C work together for 3 days.

[0105] It can be seen that for different production plan requirements, the trained production management model can output different production plans by combining employee data and production data, and each output production plan can meet the user's requirements, thereby improving the scheduling optimization in production management.

[0106] In an embodiment of the present application, S400 includes:

[0107] S410, collect production demand data and input the production demand data into the trained production management model;

[0108] S420. Obtain the production plan output by the trained production management model;

[0109] S430. Perform production scheduling according to the production plan.

[0110] It should be noted that after the production control model is trained, only the production demand data needs to be collected and input into the trained production management model to obtain the corresponding production plan, and then production can be arranged according to this production plan.

[0111] In an embodiment of the present application, the production demand data includes one or more of production cycle, production model productivity, and production cost.

[0112] It should be noted that since the material cost is fixed when the delivery quantity is certain, the production cost is mainly used to adjust the number of employees.

[0113] In an embodiment of the present application, the production model includes a production equipment model, a production line model composed of multiple production equipment models, and a factory model composed of multiple production line models.

[0114] It should be noted that the productivity of the production model calculated by Formula 1 is actually the total number of products that the entire factory model can produce within time t. Since each factory model contains multiple production lines and the production lines cooperate with each other, the productivity of the factory model is limited by the production line with the lowest production efficiency. Similarly, since each production line model contains multiple production equipment models and the production equipment models cooperate with each other, the productivity of the production line is limited by the production equipment model with the lowest productivity. Therefore, when calculating the productivity of the entire factory model, it is necessary to use the production equipment model with the lowest productivity as the basis.

[0115] The technical features of the above-described embodiments can be combined arbitrarily, and there is no limitation on the execution order of the method steps. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as these combinations of technical features do not conflict, they should be considered to be within the scope described in this specification.

[0116] The above-described embodiments only represent several implementation manners of the present application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. An intelligent production scheduling optimization system based on digital twin, characterized in that, Including: Production module; Data acquisition module, which acquires the basic data of the production module through the data acquisition module; Management module, the management module is communicatively connected to the data acquisition module, and realizes the production plan management of the production module through the management module.

2. The intelligent production scheduling optimization system based on digital twin according to claim 1, characterized in that, The basic data includes equipment data, production data and employee data.

3. An intelligent production scheduling optimization method based on digital twin, applied to the intelligent production scheduling optimization system according to claims 1 and 2, characterized in that, Including: Collect multiple pieces of basic data, preprocess the basic data to obtain preprocessed data; Create a production management model; Input the preprocessed data into the production management model, so that the production management model continuously learns the relationship between production data and employee data, and obtains a trained production management model; Collect production data requirements, and obtain a production plan based on the production data requirements.

4. The intelligent production scheduling optimization method based on digital twin according to claim 3, wherein Collect multiple pieces of basic data, preprocess the basic data to obtain preprocessed data, including: Create a basic database; Set collection parameters; the collection parameters include collection location, collection frequency and collection period; Based on the collection parameters, collect equipment data, production data and employee data respectively, and put all the collected data into the basic database; the equipment data includes the production data includes temperature data and energy consumption data, and the employee data includes employee efficiency; Randomly select a piece of basic data from the basic database; Judge whether there is missing data in the basic data. If there is missing data, fill in the data; Return to randomly select a piece of basic data from the basic database until all the basic data in the basic database have been selected, and obtain preprocessed data.

5. The intelligent production scheduling optimization method based on digital twin according to claim 4, wherein Input the preprocessed data into the production management model, so that the production management model continuously learns the relationship between production data and employee data, and obtains a trained production management model, including: Divide all the preprocessed data into a training set and a test set according to a random ratio; Input the training set into the production management model to train the production management model until a trained production management model is obtained; the trained production management model has the ability to automatically output the corresponding production plan according to the input production demand data; Input the test set into the trained production management model to verify whether the trained production management model is trained.

6. The intelligent production scheduling optimization method based on digital twin according to claim 5, wherein Input the training set into the production management model to train the production management model until a trained production management model is obtained, including: Randomly select a piece of basic data from the training set based on the collection location; Input the equipment data into the production management model, and construct a production model through the production management model; Establish the coupling relationship between production data and employee data, and use the coupling relationship of production data-employee data as a training sample; Return to randomly select a piece of basic data from the training set based on the collection location until all the basic data in the training set have been selected, and obtain multiple training samples.

7. The intelligent production scheduling optimization method based on digital twin according to claim 6, wherein Input the training set into the production management model to train the production management model until a trained production management model is obtained, and also includes: Select a training sample and input the training sample into the production management model; Calculate the productivity of each production model through Formula 1; Among them, C is the productivity of the production model, Q i is the production efficiency of the i-th production equipment model, t is the working time of the i-th production equipment model, and k is the total number of production equipment models; Statistically calculate the production efficiency of each employee; The productivity based on the production model and the production efficiency of employees construct a production evaluation function through Formula 2; wherein, R is the production evaluation function, T is the total production time, N is the productivity of the model, V is the production cost, α is the evaluation coefficient of the total production time, β is the evaluation coefficient of the productivity of the model, and γ is the evaluation coefficient of the production cost, and α + β + γ = 1.

8. The intelligent production scheduling optimization method based on digital twin according to claim 7, characterized in that Collect production data requirements and obtain a production plan based on the production data requirements, including: Collect production demand data and input the production demand data into the trained production management model; Obtain the production plan output by the trained production management model; Conduct production scheduling according to the production plan.

9. The intelligent production scheduling optimization method based on digital twin according to claim 8, characterized in that The production demand data includes one or more of a production cycle, the productivity of a production model, and a production cost.

10. The intelligent production scheduling optimization method based on digital twin according to claim 9, wherein, The production model includes a production equipment model, a production line model composed of multiple production equipment models, and a factory model composed of multiple production line models.

Images