Battery bracket workshop logistics planning method and system based on digital twinning

Abstract

The invention discloses a battery bracket workshop logistics planning method and system based on digital twinning. The method comprises the steps of production line layout design, workshop digital twinning scene model construction, multi-difference factor analysis, deep learning model establishment, planning scheme iteration, principal component analysis data dimensionality reduction and productivity prediction and evaluation. According to the method, result evaluation is quantitatively performed on the production model through digital twinborn simulation of the production line and deep iterative analysis of data, so that the workshop is optimized and improved under the condition of minimum capital investment, more accurate investment budget is achieved, and the maximum economic benefit is obtained with the minimum capital investment.

Description

Technical field:

[0001] The patent of this invention belongs to the field of logistics engineering and intelligent manufacturing, and specifically relates to a method and system for logistics planning of a battery bracket workshop based on digital twins. Background technology:

[0002] Under the background of the "dual carbon" strategy, the automotive industry is collectively embracing electrification, and the order demand for the new energy vehicle supporting parts industry is expanding rapidly. In order to meet the order demand, the expansion of production lines is imperative. On the one hand, the traditional workshop design is based on assets for layout, starting with equipment and tooling, and finally considering the flow of processes. Improving the layout on this basis is often only a partial logistics improvement. Based on the simulation and layout of the workshop with digital twins, we can start from the order demand, use a pull-type method and means, and optimize the workshop production process parameters to minimize the waste in the operation process, the fastest logistics speed, the shortest logistics distance, the least work-in-progress, and the maximum area utilization rate, thereby improving the operation efficiency. On the other hand, the development and application of deep learning models have become mature. By constructing a deep learning model of production capacity and workshop influencing factors, the impact of various workshop influencing factors on production capacity and beat is described in a quantitative way, and the future order demand and production capacity beat are predicted by model, so that under the limited capital budget, a more accurate investment budget can be achieved for workshop optimization and improvement, and the minimum capital investment can be used to exchange for the maximum capacity improvement. Summary of the invention:

[0003] The technical problem to be solved by the present invention is: to provide a method and system for logistics planning of a battery bracket workshop based on digital twins, including twin simulation module design, correlation analysis module design, twin optimization module design, iterative optimization module design, and data fitting module design, which can achieve early optimization and verification of the production capacity and logistics scheduling of new energy power battery bracket workshops to a certain extent.

[0004] The present invention solves the technical problem by adopting the following technical solution: a method and system for logistics planning of a battery bracket workshop based on digital twin, comprising the following steps:

[0005] Step (1), design the production line layout according to the planned cycle time and process route of the workshop;

[0006] Step (2), constructing a workshop digital twin scenario model and inputting data for simulation operation;

[0007] Step (3), based on multi-factor variance analysis, analyze the impact of multiple control variables on the observed variables;

[0008] Step (4), constructing a production line layout evaluation criterion for the digital twin scenario model of the new energy power battery bracket workshop, building a deep learning framework, setting an optimization function, and quantitatively evaluating the simulation results of the design scheme;

[0009] Step (5), iteratively optimizing the simulation results based on the evaluation criteria and designing an iterative optimization module;

[0010] Step (6) constructs a mechanism model to describe the factors affecting output and workshop, uses principal component analysis to reduce the dimension of the data, gives an expression between output and production factors, and predicts and evaluates production capacity.

[0011] Furthermore, the step (1) designs the production line layout according to the planned cycle time and process route of the workshop, specifically including:

[0012] (1.1) According to the planned production capacity requirements, calculate the production line rhythm and staffing requirements, and design the production line plan based on the one-flow principle, combined with the process requirements and the workshop environment;

[0013] (1.2) Calculate the size and configuration of logistics equipment (pallets, AGVs, transport vehicles, etc.) based on the size and weight of the products produced;

[0014] Furthermore, the step (2) constructs a workshop digital twin scenario model and inputs data for simulation operation; the specific implementation is as follows:

[0015] (2.1) Construct a digital twin simulation model scenario for new energy power battery brackets. The scenario covers welding robots, slide fixtures, CNC machine tools, AGV transfer carts, friction welding fixtures, cleaning lines, visual inspection systems, airtight testing machines, and helium leak detection stations in the new energy power battery bracket workshop, truly restoring the current status of workshop production and logistics operations;

[0016] (2.2) Filter the 6-month real production history data in the MES system and the production line layout and logistics scheduling strategy planned and designed in step (1) as input to drive the simulation operation of the digital twin scenario of the new energy power battery planning bracket workshop constructed in step (2.1);

[0017] Furthermore, the step (3) is based on multi-factor variance analysis to analyze the impact of multiple control variables on the observed variables; the specific implementation is as follows:

[0018] (3.1) The evaluation index of production line output capacity is affected by multiple independent variable factors, such as equipment failure rate, buffer size configuration, AGV transport vehicle configuration, station personnel configuration, supplier arrival punctuality rate, equipment failure repair time, product changeover time. Determine the range of independent variables that affect the evaluation index of production line output capacity;

[0019] (3.2) The fluctuations in the production process of the production line conform to the normal distribution. The multi-factor variance analysis method is used to calculate and analyze the impact of multiple control variables on the observed variables, and the variables with correlation coefficients <0.3 are screened. The variables with correlation coefficients >0.7 are simulated, analyzed and optimized.

[0020] Furthermore, in step (4), a production line layout evaluation criterion for the digital twin scenario model of the new energy power battery bracket workshop is constructed, a deep learning framework is built, an optimization function is set, and a simulation result of the design scheme is quantitatively evaluated;

[0021] (4.1) Use the improved facility layout planning method (SLP) to plan the overall functional area layout of the workshop, and perform equipment analysis, logistics flow analysis, area demand analysis, and historical peak capacity analysis and calculation;

[0022] (4.2) Calculate the optimization function F, which is defined as

[0023] F=∑M k *W ij +ΣH k *D ij +ΣK k *Z ij +ΣN k *W k

[0024] Where M k , H k is the functional area, W ij is the raw material logistics intensity per unit area, D ij is the logistics intensity per unit area of ​​finished products, N k is the influencing factor with a large correlation coefficient calculated in step (3), W k is a weight adjustment factor; further, in step (5), the simulation results are iteratively optimized based on the evaluation criteria, and an iterative optimization module is designed;

[0025] (5.1) Define the loss function as the root mean square loss function (MSE), and use the stochastic gradient descent method to iteratively optimize the loss;

[0026] (5.2) Randomly initialize and iteratively optimize the loss function. When the loss function converges, stop the iteration and statistically observe the weights and parameter values ​​of each influencing factor.

[0027] Furthermore, in step (6), a mechanism model describing the factors affecting the output and the workshop is constructed, and principal component analysis is used to reduce the dimension of the data, and an expression between the output and the production factors is given, and the production capacity is predicted and evaluated. The specific contents are as follows:

[0028] (6.1) Through the combination of production system simulation and deep learning, a nonlinear parameter model describing the production capacity and production influencing factors of the new energy battery bracket workshop is built to quantitatively describe the complex production system process;

[0029] (6.2) Optimize the data model and use principal component analysis (PCA) to reduce the dimension of the complex production system data model, retain the main data features, and conduct capacity prediction and evaluation;

[0030] On the other hand, the present invention proposes a digital twin new energy power battery bracket workshop logistics planning and optimization system, including:

[0031] Twin simulation module: used to build a production simulation twin scenario of a new energy power battery bracket workshop, input simulation data and run, including: building a digital twin scenario data model of a battery bracket workshop, the scenario covers production equipment, logistics equipment, storage equipment, inspection and testing equipment, personnel factors, etc. in the production process. Describe and depict the production status of the real power battery bracket workshop, use real historical production data, logistics data, equipment operation and maintenance data and product quality data as strategy input, and drive the operation of the digital twin scenario of the power battery bracket workshop;

[0032] The correlation analysis module analyzes the impact of multiple control variables on observed variables based on multi-factor variance analysis. Specifically, it includes: determining the input range of independent variables that affect the production line output capacity evaluation index; using multi-factor variance analysis to calculate and analyze the impact of multiple control variables on observed variables, and screening variables with correlation coefficients > 0.7 for key simulation analysis and optimization;

[0033] The twin optimization module is used to construct the evaluation criteria for the battery bracket workshop production system, construct a mathematical mechanism model, and evaluate and optimize the simulation results; specifically, it includes: building a power battery bracket workshop production status evaluation mechanism model based on a deep learning framework; the evaluation criteria are driven by process expert experience, lean management experience, production management experience, and historical operation and maintenance, and use the improved facility layout planning method (SLP) to plan the overall functional area layout of the workshop, and perform equipment analysis, logistics flow analysis, area demand analysis, historical peak capacity analysis and other calculations. Calculate the optimization function F, evaluate and optimize the simulation operation results based on the evaluation criteria, and the optimization function is defined as

[0034] F=∑M k *W ij +ΣH k *D ij +ΣK k *Z ij +ΣN k *W k

[0035] Where Mk , H k is the functional area, W ij is the raw material logistics intensity per unit area, D ij is the logistics intensity per unit area of ​​finished products, N k is the influencing factor with a large correlation coefficient calculated in step (3), W k is the weight adjustment factor;

[0036] An iterative optimization module performs iterative optimization on the production line mechanism model; further, when the output simulation operation result does not meet the evaluation criteria, the iterative optimization module sets the number of iterative optimization times, and when the number of iterative optimization times is less than the set value, the relevant parameters are adjusted and the operation is performed again, and when the output simulation operation result meets the evaluation criteria, the iterative operation is terminated;

[0037] The data fitting module builds a mechanism model that describes the workshop production rhythm, output and production influencing factors, uses principal component analysis (PCA) to reduce the data dimension, gives the relationship between the workshop output and the influencing factors, and performs capacity prediction;

[0038] The advantages of the present invention compared with the prior art are: the traditional workshop construction method, due to insufficient estimation of enterprise development strategy and capacity planning, incomplete consideration of factors such as overall logistics planning, equipment commissioning and installation, turnover volume during the turnover period, production line layout and process design, and unreasonable design scheduling, leads to waste such as long production distance, insufficient equipment or low utilization rate, and unbalanced processes in the new workshop after it is put into production. The present invention is based on digital twin workshop simulation and layout. By constructing a deep learning model, the workshop production model is analyzed and modeled, the workshop production process parameters are optimized, and the future order demand and production capacity rhythm are modeled to minimize waste in the operation process, with the fastest logistics speed, the shortest logistics distance, the least work-in-progress, and the maximum area utilization rate, improving operation efficiency, and achieving a more accurate investment budget for workshop optimization and improvement, with the minimum capital investment in exchange for the maximum capacity improvement. Description of the drawings:

[0039] Figure 1 This is a simulation diagram of the digital twin production system.

[0040] Figure 2 This is a structural diagram of the deep learning framework of the present invention.

[0041] Figure 3 It is the technical roadmap of the present invention. Specific implementation method:

[0042] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. It is obvious that the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the protection scope of the present invention.

[0043] The present invention relates to a method and system for logistics planning of a battery bracket workshop based on digital twins, the method comprising the following steps:

[0044] Step (1), design the production line layout according to the planned cycle time and process route of the workshop;

[0045] Step (2), constructing a workshop digital twin scenario model and inputting data for simulation operation;

[0046] Step (3), based on multi-factor variance analysis, analyze the impact of multiple control variables on the observed variables;

[0047] Step (4), constructing a production line layout evaluation criterion for the digital twin scenario model of the new energy power battery bracket workshop, building a deep learning framework, setting an optimization function, and quantitatively evaluating the simulation results of the design scheme;

[0048] Step (5), iteratively optimizing the simulation results based on the evaluation criteria and designing an iterative optimization module;

[0049] Step (6) constructs a mechanism model to describe the factors affecting output and workshop, uses principal component analysis to reduce the dimension of the data, gives an expression between output and production factors, and predicts and evaluates production capacity.

[0050] Furthermore, the step (1) designs the production line layout according to the planned cycle time and process route of the workshop, specifically including:

[0051] (1.1) Calculate the production line rhythm and staffing requirements based on the planned capacity requirements, and design the production line plan based on the one-flow principle, combined with the process requirements and the workshop environment;

[0052] (1.2) Calculate the size and configuration of logistics equipment (pallets, AGVs, transport vehicles, etc.) based on the size and weight of the products produced;

[0053] Furthermore, the step (2) constructs a workshop digital twin scenario model and inputs data for simulation operation; the specific implementation is as follows:

[0054] (2.1) Construct a digital twin simulation model scenario for the new energy power battery bracket. According to step (1.2), use the modeling software to model the data of production equipment, logistics equipment, etc. According to step (1.1), use the production system simulation software to model the production line. The scenario covers the welding robots, slide tooling, CNC machine tools, AGV transfer carts, friction welding tooling, cleaning lines, visual inspection systems, airtight testing machines, helium leak detection stations, etc. in the new energy power battery bracket workshop, truly restoring the current status of workshop production and logistics operations, such as Figure 1 As shown;

[0055] (2.2) Filter the historical production data of 6 months in the MES system, use the production line layout and logistics scheduling strategy planned and designed in step (1) as input, and drive the simulation operation of the digital twin scenario of the new energy power battery planning bracket workshop constructed in step (2.1) with data;

[0056] Furthermore, the step (3) is based on multi-factor variance analysis to analyze the impact of multiple control variables on the observed variables; the specific implementation is as follows:

[0057] (3.1) The evaluation index of production line output capacity is affected by multiple independent variable factors, such as equipment failure rate, buffer size configuration, AGV transport vehicle configuration, personnel configuration at each workstation, supplier arrival punctuality rate, equipment failure repair time, product changeover time, etc. Determine the range of independent variables for the evaluation index of production line output capacity;

[0058] (3.2) The fluctuations in the production process of the production line conform to the normal distribution. The multi-factor variance analysis method is used to calculate and analyze the impact of multiple control variables on the observed variables, and the variables with correlation coefficients <0.3 are screened. The variables with correlation coefficients >0.7 are simulated, analyzed and optimized.

[0059] ① Propose the original hypothesis

[0060] There is no significant difference in the means of the populations under different levels of the observed variables of the factor variables in (3.1) (failure rate of equipment, configuration of buffer size, configuration of AGV transport carts, configuration of personnel at each workstation, on-time arrival rate of suppliers, equipment failure repair time, equipment changeover time, etc.);

[0061] H0:μ1=μ2=μ3=μ4=μ i =......=μ k , where μ i is the mean of the i-th level;

[0062] H1:μ i (i=1,2,3,...k) are not all equal

[0063] ②Construct test statistics

[0064] ③ Observed variable variance decomposition

[0065] Total error sum of squares

[0066] A factor error sum of squares

[0067] B factor error sum of squares

[0068] Interaction sum of squares

[0069] The sum of squares of the error terms SSE = SST-SSA-SSB-SSAB

[0070]

[0071] In the formula, SST is the total variance of the observed variable, SSA and SSB are the variance caused by the independent effects of the control variables A and B, SSAB is the variance caused by the interaction of the control variables A and B, and SSE is the variance caused by random factors. SSA+SSB is the main effect, and SSAB is the multiple interaction effect;

[0072] ④ Correlation coefficient calculation

[0073]

[0074] In the formula, R is the correlation coefficient, SSAB is the multi-factor interaction effect, SST is the total variance of the observed variable, R<0.3, the independent variable and the dependent variable are weakly correlated, R>0.7, the independent variable and the dependent variable are strongly correlated. Factors with R>0.7 were selected for further analysis and measurement;

[0075] Furthermore, in step (4), the production line layout evaluation criteria of the digital twin scenario model of the new energy power battery bracket workshop are constructed, a deep learning framework is built, an optimization function is set, and the simulation results of the design scheme are quantitatively evaluated, such as Figure 2 As shown;

[0076] (4.1) Use the improved facility layout planning method (SLP) to plan the overall functional area layout of the workshop, and perform calculations such as equipment analysis, logistics flow analysis, area demand analysis, and historical peak capacity analysis;

[0077] (4.2) Calculate the optimization function L, which is defined as

[0078] L=ΣM k *W ij +ΣH k *D ij +ΣK k*Z ij +ΣN k *W k

[0079] Where M k , H k is the functional area, W ij is the raw material logistics intensity per unit area, D ij is the logistics intensity per unit area of ​​finished products, N k is the influencing factor with a large correlation coefficient calculated in step (3), W k is the weight adjustment factor;

[0080] Furthermore, in step (5), the simulation results are iteratively optimized based on the evaluation criteria, and an iterative optimization module is designed;

[0081] (5.1) Define the loss function as the root mean square loss function and use the stochastic gradient descent method to iteratively optimize the loss;

[0082]

[0083] In the formula, is the simulated observation value, L i Calculate the value of the optimization function in (4.2);

[0084] (5.2) Define the stochastic gradient descent method to iteratively optimize the loss, and its iterative formula is:

[0085]

[0086] In the formula, η represents the learning rate, represents the gradient of the loss function J(θ), x (i) ;y (i) is the training sample, θ 1 represents the updated parameter value, θ 0 Represents the current parameter value. The loss function is randomly initialized and iteratively optimized. When the loss function converges, the iteration is stopped, and the weights and parameter values ​​of each influencing factor are statistically observed;

[0087] Furthermore, in step (6), a mechanism model describing the factors affecting the output and the workshop is constructed, and principal component analysis is used to reduce the dimension of the data, and an expression between the output and the production factors is given, and the production capacity is predicted and evaluated. The specific contents are as follows:

[0088] (6.1) By combining production system simulation and deep learning, a nonlinear parameter model describing the production capacity and influencing factors of the new energy battery bracket workshop is built according to steps (4) and (5) to quantitatively describe the complex production process;

[0089] (6.2) Further reduce the dimension of the complex data model, optimize the data model, use principal component analysis (PCA) to reduce the dimension of the simulation data model, retain the main data features, and perform capacity prediction and evaluation; the specific operations are as follows:

[0090] ①Decentralize the feature matrix to obtain a new matrix X

[0091]

[0092] ② Calculate the covariance moment C of the decentralized matrix image

[0093]

[0094] ③ Perform eigendecomposition on the covariance matrix C and find the eigenvalue λ of the covariance matrix k , and the corresponding eigenvector v k , that is, Cv k =λ k v k

[0095] ④ Arrange the eigenvectors in descending order from left to right according to the corresponding eigenvalues, and take the first k columns to form the matrix W;

[0096] ⑤ Calculate the sample features after the feature vector is reduced to k dimensions through Y=XW.

[0097] The system structure diagram of the present invention is as follows: Figure 3 As shown, the specific implementation is as follows:

[0098] Figure 3 1 in represents the twin simulation module, Figure 3 The 2 in the figure indicates the relevant analysis module. Figure 3 The 3 in the figure represents the twin optimization module. Figure 3 The 4 in represents the iterative optimization module; Figure 3 The 5 in represents the data fitting module;

[0099] The twin simulation module is specifically implemented as follows:

[0100] ① Build a digital twin scenario data model for the battery bracket workshop, which covers production equipment, logistics equipment, warehousing equipment, inspection and testing equipment, and personnel factors in the production process;

[0101] ② Describe and depict the production status of the real power battery bracket workshop, use real historical production data, logistics data, equipment operation and maintenance data, and product quality data as strategy inputs to drive the operation of the digital twin scene of the power battery bracket workshop;

[0102] The related analysis module is specifically implemented as follows:

[0103] ① Based on the data collection factors in (2), determine the input range of the independent variables that affect the production line output capacity evaluation index;

[0104] ② Use multi-factor variance analysis to calculate and analyze the impact of multiple control variables on observed variables, screen out variables with low correlation, and focus on the impact of variables with correlation coefficients > 0.7;

[0105] The twin optimization module is specifically implemented as follows:

[0106] ①Build a production status evaluation mechanism model for power battery bracket workshop based on deep learning framework;

[0107] ② The evaluation criteria are driven by process expert experience, lean management experience, production management experience, and historical operation and maintenance. The improved facility layout planning method (SLP) is used to plan the overall functional area layout of the workshop, and to perform equipment analysis, logistics flow analysis, area demand analysis, and historical peak capacity analysis and calculation;

[0108] The iterative optimization module is specifically implemented as follows:

[0109] ① Iteratively optimize the production line mechanism model based on the deep learning framework, loss function and stochastic gradient descent method built in (4);

[0110] ② Further, when the output simulation running result does not meet the evaluation criteria (4), the iterative optimization module sets the number of iterative optimization times. When the number of iterative optimization times is less than the set value, the relevant parameters are adjusted and the process is run again. When the output simulation running result meets the evaluation criteria, the iterative operation is terminated.

[0111] The data fitting module is specifically implemented as follows:

[0112] According to (5), a mechanism model describing the relationship between workshop output and production influencing factors is constructed, and principal component analysis (PCA) is used to reduce the data dimension to perform capacity prediction and evaluation;

[0113] In summary, the present invention uses digital twin production system simulation methods and combines deep learning framework to describe the production mechanism model of the power battery bracket production line, which can realize the advance optimization and verification of battery bracket workshop logistics to a certain extent. Under the limited capital budget, it can achieve a more accurate investment budget for workshop optimization and improvement, and exchange the minimum capital investment for the maximum capacity improvement.

[0114] The contents not described in detail in the specification of the present invention belong to the prior art known to the professional and technical personnel in this field.

[0115] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method and system for logistics planning of battery rack workshop based on digital twin, characterized in that: The steps include: Step (1), design the production line layout according to the planned cycle time and process route of the workshop; Step (2), constructing a workshop digital twin scenario model and inputting data for simulation operation; Step (3), based on multi-factor variance analysis, analyze the impact of multiple control variables on the observed variables; Step (4): Based on the workshop digital twin scenario model, construct the production line layout evaluation criteria, build a deep learning framework, iterate the design scheme, and conduct quantitative evaluation of the results; Step (5), iteratively optimizing the simulation results based on the evaluation criteria and designing an iterative optimization module; Step (6) constructs a mechanism model to describe the factors affecting output and workshop, uses principal component analysis to reduce the dimension of the data, gives an expression between output and production factors, and predicts and evaluates production capacity.

2. According to claim 1, a method and system for battery bracket workshop logistics planning based on digital twins is characterized in that: The step (1) designs the production line layout according to the planned cycle time and process route of the workshop, specifically including: (1.1) According to the planned production capacity requirements, calculate the production line rhythm and staffing requirements, and design the production line plan based on the one-flow principle, combined with the process requirements and the workshop environment; (1.2) Calculate the size and configuration of logistics equipment (pallets, AGVs, transport vehicles, etc.) based on the size and weight of the products.

3. According to claim 1, a method and system for battery bracket workshop logistics planning based on digital twins is characterized in that: The step (2) constructs a workshop digital twin scenario model and inputs data for simulation operation; (2.1) Construct a digital twin simulation model scenario for new energy power battery brackets. The scenario covers the welding robots, slide fixtures, CNC machine tools, AGV transfer carts, friction welding fixtures, cleaning lines, visual inspection systems, airtight testing machines, and helium leak detection stations in the new energy power battery bracket workshop, truly restoring the current status of workshop production and logistics operations; (2.2) The real production history data of 6 months in the MES system and the production line layout and logistics scheduling strategy planned and designed in step (1) are filtered and used as input to drive the simulation operation of the digital twin scenario of the new energy power battery planning bracket workshop constructed in step (2.1).

4. According to claim 1, a method and system for battery bracket workshop logistics planning based on digital twins is characterized in that: The step (3) is to analyze the influence of multiple control variables on the observed variables based on multi-factor variance analysis; (3.1) The evaluation index of production line output capacity is affected by multiple independent variable factors, such as equipment failure rate, buffer size configuration, AGV transport vehicle configuration, personnel configuration at each workstation, supplier arrival punctuality rate, equipment failure repair time, and equipment changeover time. Determine the range of independent variables that affect the evaluation index of production line output capacity; (3.2) The fluctuations in the production process of the production line conform to the normal distribution. The multi-factor variance analysis method is used to calculate and analyze the impact of multiple control variables on the observed variables, and the variables with correlation coefficients <0.3 are screened. The variables with correlation coefficients >0.7 are simulated, analyzed and optimized.

5. According to claim 1, a method and system for battery bracket workshop logistics planning based on digital twins is characterized in that: In step (4), based on the workshop digital twin scenario model, a production line layout evaluation criterion is constructed, a deep learning framework is built, the design scheme is iterated, and the results are quantitatively evaluated; (4.1) Use the improved facility layout planning method (SLP) to plan the overall functional area layout of the workshop, and perform equipment analysis, logistics flow analysis, area demand analysis, and historical peak capacity analysis and calculation; (4.2) Define and calculate the optimization function L.

6. A method and system for battery rack workshop logistics planning based on digital twin according to claim 1, characterized in that: The step (5) sets the loss function and the optimizer, and iteratively optimizes the simulation results based on the evaluation criteria; (5.1) The loss function selects the root mean square loss function and uses the stochastic gradient descent method to iteratively optimize the loss; (5.2) The loss function is randomly initialized and iteratively optimized. When the loss function tends to converge, the iteration is stopped, and the weights and parameter values ​​of each influencing factor are statistically observed.

7. A method and system for battery rack workshop logistics planning based on digital twin according to claim 1, characterized in that: The step (6) constructs a mechanism model describing the factors affecting the output and the workshop, uses principal component analysis (PCA) to reduce the dimension of the data, gives an expression between the factors affecting the workshop output and the influencing factors, and performs capacity prediction; (6.1) The deep learning method provides a nonlinear parameter model that describes the factors affecting production and workshops. Complex data models need to be reduced in dimension to optimize the data model. (6.2) Use principal component analysis (PCA) to reduce the dimension of the factory simulation data model, further optimize the production line production data model, retain the main feature parts, and perform capacity prediction and evaluation.

Images