An automobile accessory die casting mold manufacturing cost prediction method

Abstract

The application discloses a kind of automobile accessory die casting mould manufacturing cost prediction method, belong to automobile accessory manufacturing and digitization field, solve the low precision of traditional cost prediction, not cover full life cycle, cannot dynamic iteration and other industry pain points.The method fuses soft power industry metaverse, digital twin and industrial big data technology, successively completes die casting mould 1:1 digital twin body construction and full life cycle basic data acquisition, the building of multi-dimensional decomposition model of five core cost dimensions, the training of AI intelligent cost prediction model fusing LSTM and random forest algorithm, the construction of dynamic iteration optimization mechanism containing four links, finally output cost prediction result and multidimensional optimization suggestions such as material, process.The method relies on soft power multi-digitalization platform and module implementation, can accurately calculate each dimension cost and weighted summation, combine simulation deduction to correct error to make error≤5%, and can also update and multi-end feedback continuously optimize model through real-time data.

Description

Technical Field

[0001] This invention relates to the fields of automotive parts manufacturing and digital technology, and in particular to a method for predicting the manufacturing cost of automotive parts die-casting molds. Background Technology

[0002] With the rapid development of the new energy vehicle industry, auto parts are transforming towards lightweight, precision, and integration. As the core equipment for auto parts manufacturing, the manufacturing cost of die-casting molds directly determines the production cost and market competitiveness of auto parts.

[0003] There are several pain points in the current cost prediction of automotive parts die-casting mold manufacturing: First, traditional prediction methods rely on manual experience and are based on macroscopic parameter estimation, which is highly subjective and has low accuracy, making it difficult to adapt to the cost breakdown requirements of complex die-casting molds, with prediction errors generally exceeding 15%; Second, existing prediction models mostly focus on a single manufacturing stage, failing to consider the cost correlation of the entire life cycle of mold design, processing, trial molding, and operation and maintenance, and cannot dynamically respond to real-time variables such as raw material price fluctuations, equipment wear and tear, and process optimization; Third, there is a lack of deep integration with industrial digital platforms, making it impossible to use historical production data and industry benchmark data for self-iterative optimization, and thus failing to meet the demand for rapid cost prediction in the mass customization production of automotive parts.

[0004] As a leading industrial internet service provider in China, iSoftStone possesses profound technological expertise and rich industry experience in areas such as industrial metaverse, digital twins, and industrial big data. Its iSSMeta digital simulation and deduction platform has been widely applied in discrete manufacturing and process manufacturing, providing manufacturing enterprises with full-stack digital solutions. Currently, there is no technical solution that deeply integrates iSoftStone's core digital technologies with cost prediction for automotive parts die-casting molds. Existing prediction methods cannot fully utilize the simulation, data aggregation, and intelligent analysis capabilities of iSoftStone's industrial digital platform, making it difficult to address the core pain points of low accuracy, low efficiency, and poor dynamism in industry cost prediction. Summary of the Invention

[0005] In order to solve the problems in the background art, the present invention proposes a method for predicting the manufacturing cost of die-casting molds for automotive parts.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for predicting the manufacturing cost of automotive parts die-casting molds, incorporating iSoftStone's industrial metaverse, digital twin, and industrial big data technologies, includes the following specific steps: Step 1: Construct a digital twin of the iSSMeta die-casting mold and collect basic data throughout the entire lifecycle; Relying on the iSSMeta digital simulation and deduction platform of iSoftStone, construct a 1:1 digital twin of the die-casting mold of automotive parts, collect basic data of each stage of the entire lifecycle, and store the data on the iSoftStone industrial cloud platform after iSoftStone data cleaning and standardization. Step 2: Construct a multi-dimensional cost decomposition model based on iSoftStone's industrial big data capabilities; break down mold costs into five core dimensions, use dynamic calculation methods, and combine industry benchmark data to achieve refined decomposition. Step 3: Construct the iSoftStone AI intelligent cost prediction model to achieve accurate predictions; relying on the iSoftStone AI algorithm platform, integrating LSTM and random forest algorithms, training the model using a standardized dataset, predicting costs in each dimension and summing them with weights, and using the iSSMeta platform to correct errors, ensuring that the error is ≤5%; Step 4: Build a dynamic iterative optimization mechanism to improve predictive adaptability; rely on real-time data updates from iSoftStone Industrial Cloud, combine actual and predicted cost errors to iteratively optimize the AI ​​model, and achieve multi-terminal feedback through the iSoftStone collaborative management platform to adapt to industry changes; Step 5: Output cost forecast results and optimization suggestions; Output cost forecast results and their composition through the iSoftStone visualization module, and provide multi-dimensional cost optimization suggestions in conjunction with iSoftStone's digital solutions.

[0007] Preferably, the basic data includes design data, material data, processing data, trial molding and maintenance data, and industry benchmark data; wherein the design data is imported through the iSoftStone CAD collaborative design module, and uses the improved BREPNet architecture and MoE hybrid expert model to automatically identify key mold structures and quantify design complexity scores; material data is collected in real time from upstream suppliers through the iSoftStone Industrial Big Data Platform; processing data is collected in real time from processing equipment through the iSoftStone AIoT module; trial molding and maintenance data is collected through digital twin simulation and maintenance system integration; and industry benchmark data is formed by aggregating historical industry data through the iSoftStone Industrial Big Data Platform.

[0008] Preferably, the dynamic calculation method is as follows: Design cost = Design complexity score × Design cost per unit complexity + Design iteration cost. The design cost per unit complexity is based on data from the Softcom industry benchmark database and is dynamically adjusted in conjunction with the design cycle and designers' salaries. Material cost = (mold material usage × (1 + loss rate) × real-time material unit price) + transportation cost + warehousing cost. Mold material usage is automatically calculated through a digital twin 3D model, and the loss rate is dynamically optimized based on historical data from the iSoftStone Industrial Big Data Platform. Processing cost = labor cost + equipment depreciation cost + energy cost + tooling cost. Labor cost is dynamically calculated by simulating the processing process through a digital twin. Equipment depreciation and energy cost are calculated based on real-time data collected by the Softcom AIoT module. Trial molding and testing costs = trial molding material loss + trial molding equipment energy consumption + testing fees + scrap disposal costs. The number of trial molding tests and losses can be predicted by simulating the trial molding process using a digital twin. Operation and maintenance costs and hidden costs = maintenance costs + fault repair costs + management costs + environmental compliance costs. Maintenance costs are based on the mold's lifespan and maintenance cycle prediction, while environmental compliance costs are quantified through the Softcom dual-carbon control module.

[0009] Preferably, the standardized dataset training model process includes: dividing the standardized dataset into a training set, a validation set, and a test set; training the model using the training set; adjusting the model parameters using the validation set; calculating the prediction error using the test set; and correcting the error using digital twin simulation capabilities; the weights of each dimension of cost are obtained through training with historical data from the Softcom Industrial Big Data Platform.

[0010] Preferably, the dynamic iterative optimization mechanism includes four stages: real-time data update, model iterative optimization, scenario adaptive adjustment, and multi-terminal collaborative feedback. Real-time data update synchronously collects information on changes in raw material prices, equipment status, and industry benchmark data. Model iterative optimization uses the error data between actual and predicted costs to adjust model parameters and cost weights. Scenario adaptive adjustment uses a digital twin platform to simulate new manufacturing scenarios and optimize model parameters. Multi-terminal collaborative feedback enables feedback and interaction among personnel at each stage through the Softcom collaborative management platform.

[0011] Preferably, the multi-dimensional cost optimization suggestions include material optimization, process optimization, design optimization, and operation and maintenance optimization. Among them, material optimization recommends the most cost-effective material substitution and procurement scheme, process optimization recommends the optimal processing technology route, design optimization recommends the simplification of mold structure and reduction of design complexity, and operation and maintenance optimization recommends reasonable maintenance cycle and maintenance scheme.

[0012] Preferably, the cost prediction method is implemented based on the iSSMeta digital simulation platform, industrial big data platform, AI algorithm platform, AIoT module, MES production execution platform, collaborative management platform, and dual carbon control module of iSoftStone, and can be seamlessly integrated with the existing production system of auto parts manufacturing enterprises.

[0013] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: This invention, through the integrated application of Softcom digital twin, AI algorithm and big data analysis, controls the prediction error to within 1%, which is far better than the error level of more than 15% of the existing technology. It achieves accurate prediction of the manufacturing cost of die casting molds and provides reliable support for enterprise cost control and pricing decisions. This invention relies on the real-time data update capability and dynamic iterative optimization mechanism of the Softcom Industrial Cloud Platform to respond in real time to variables such as raw material price fluctuations, equipment status changes, process optimization, and industry technology iterations, ensuring the timeliness and accuracy of prediction results and adapting to different types and complexities of automotive parts die-casting molds. This invention helps enterprises identify key aspects of cost control through multi-dimensional cost decomposition and targeted optimization suggestions, effectively reducing manufacturing costs by 1.9%-5.6%, reducing material waste, mold trial losses, equipment idleness, and other problems, thereby improving the economic benefits of enterprises. This invention integrates Softcom's core digital technologies to achieve the integration, analysis, and visualization of data throughout the entire lifecycle of molds. It breaks down data silos at each stage, improves the digitalization and intelligence of enterprise production and operation, and can be seamlessly connected with the enterprise's existing production system, reducing implementation costs. Attached Figure Description

[0014] Figure 1 A flowchart is shown for a method for predicting the manufacturing cost of automotive parts die-casting molds according to an embodiment of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1 This embodiment focuses on the die-casting mold for an integrated die-cast body structural component of a new energy vehicle (material: aluminum alloy, weight: 12kg). This mold has a complex structure, including conformal cooling channels, large cavities, and other key features. It requires a dimensional accuracy of ≤±0.02mm and a service life of ≥100,000 mold cycles. Accurate prediction of its entire lifecycle manufacturing cost is necessary to support enterprise pricing and cost control. This embodiment relies on iSoftStone's existing digital platform, with all data sourced from actual production scenarios and iSoftStone's industry benchmark database.

[0017] Please see Figure 1 The present invention provides a technical solution: A method for predicting the manufacturing cost of automotive parts die-casting molds, incorporating iSoftStone's industrial metaverse, digital twin, and industrial big data technologies, includes the following specific steps: Step 1: Construct a digital twin of the iSSMeta die-casting mold and collect basic data throughout its entire lifecycle; relying on the iSSMeta digital simulation platform of iSoftStone, import the 3D model of the integrated die-casting mold for the body structure component, construct a 1:1 digital twin of the automotive parts die-casting mold, integrate basic data from all aspects of mold design, material procurement, precision machining, heat treatment, mold trial verification, operation and maintenance, and through iSoftStone's data cleaning and standardization module, form a standardized basic dataset for cost prediction, which is stored on the iSoftStone Industrial Cloud Platform; The basic data includes design data, material data, processing data, mold trial and maintenance data, and industry benchmark data. Design data is imported through the iSoftStone CAD collaborative design module, utilizing an improved BREPNet architecture and a MoE hybrid expert model to automatically identify key mold structures and quantify design complexity scores. Material data is collected in real-time from upstream suppliers via the iSoftStone Industrial Big Data Platform. Processing data is collected in real-time from processing equipment via the iSoftStone AIoT module. Mold trial and maintenance data is collected through digital twin simulation and integration with the maintenance system. Industry benchmark data is generated by aggregating historical industry data through the iSoftStone Industrial Big Data Platform. Specifically: Design data: The mold 3D model is imported through the Softcom CAD collaborative design module. Using the improved BREPNet architecture and MoE hybrid expert model, key structures such as mold conformal cooling channels and large cavities are automatically identified. The quantitative design complexity score is 8.5 (out of 10). The design cycle is 15 days, the designer's salary is 300 yuan / day, and the design is iterated once. Material data: Through the Softcom Industrial Big Data Platform, we connected with H13 hot work die steel suppliers and collected real-time material unit price data of 18,600 yuan / ton, purchase batch of 5 tons, transportation cost of 800 yuan, and loss rate of 3%; the die material usage was automatically calculated to be 4.8 tons through digital twin 3D model. Processing data: Through the Softcom AIoT module, the system connects to processing equipment such as five-axis machining centers, EDM machine tools, and vacuum high-pressure gas quenching equipment, and collects processing time in real time. The estimated processing time is 120 hours, equipment depreciation is 50 yuan per hour, energy consumption is 15 kWh per hour (electricity price is 0.8 yuan / kWh), and tool wear is estimated at 2000 yuan. At the same time, the system collects processing accuracy detection data synchronously through the Softcom MES manufacturing execution platform. Trial molding and maintenance data: The trial molding process is simulated using a digital twin. The number of trial molding tests is estimated at 2, with material consumption of 1500 yuan per test and testing cost of 3000 yuan. Through integration with the Softcom maintenance system, the mold maintenance cycle is estimated at 1000 cycles, with a maintenance cost of 800 yuan per cycle and a failure probability of 0.5%. Industry benchmark data: By aggregating historical cost data, industry average loss rate (2.8%), and average maintenance cost (750 yuan / time) of similar integrated die-casting molds through the Softcom Industrial Big Data Platform; All the data collected above are processed through the Softcom data cleaning and standardization module to remove abnormal data, unify formats, and align dimensions, forming a standardized cost prediction basic dataset, which is stored on the Softcom Industrial Cloud Platform.

[0018] Step 2: Construct a multi-dimensional cost decomposition model based on iSoftStone's industrial big data analysis capabilities; break down mold manufacturing costs into five core dimensions: design cost, material cost, processing cost, trial molding and testing cost, and operation and maintenance and hidden costs. The cost of each dimension is calculated dynamically and combined with iSoftStone's industry benchmark database data to achieve a refined cost decomposition. The dynamic calculation method is as follows: Design cost = Design complexity score × Design cost per unit complexity + Design iteration cost. The design cost per unit complexity is based on data from the Softcom industry benchmark database and is dynamically adjusted in conjunction with the design cycle and designers' salaries. Material cost = (mold material usage × (1 + loss rate) × real-time material unit price) + transportation cost + warehousing cost. Mold material usage is automatically calculated through a digital twin 3D model, and the loss rate is dynamically optimized based on historical data from the iSoftStone Industrial Big Data Platform. Processing cost = labor cost + equipment depreciation cost + energy cost + tooling cost. Labor cost is dynamically calculated by simulating the processing process through a digital twin. Equipment depreciation and energy cost are calculated based on real-time data collected by the Softcom AIoT module. Trial molding and testing costs = trial molding material loss + trial molding equipment energy consumption + testing fees + scrap disposal costs. The number of trial molding tests and losses can be predicted by simulating the trial molding process using a digital twin. Operation and maintenance costs and hidden costs = maintenance costs + fault repair costs + management costs + environmental compliance costs. Maintenance costs are based on the mold's lifespan and maintenance cycle prediction, while environmental compliance costs are quantified through the Softcom dual-carbon control module. Specifically: Design cost = Design complexity score × Unit complexity design cost + Design iteration cost; where the unit complexity design cost is based on similar mold data from the Softcom industry benchmark database, and is dynamically adjusted to 1200 yuan based on the design cycle and designer salaries, and the design iteration cost is 2000 yuan; the final design cost = 8.5 × 1200 + (15 × 300) + 2000 = 16700 yuan.

[0019] Material cost = (mold material usage × (1 + loss rate) × real-time material unit price) + transportation cost + warehousing cost; where warehousing cost is calculated at 0.5% of the material purchase amount, the final material cost = (4.8 × (1 + 3%) × 18600) + 800 + (4.8 × 18600 × 0.5%) = 92758.4 + 446.4 = 93204.8 yuan.

[0020] Processing cost = labor cost + equipment depreciation cost + energy cost + tooling cost; where labor cost = 120 × 300 (processing personnel salary) = 36,000 yuan, equipment depreciation cost = 120 × 50 = 6,000 yuan, energy cost = 120 × 15 × 0.8 = 1,440 yuan, tooling cost = 2,000 yuan; final processing cost = 36,000 + 6,000 + 1,440 + 2,000 = 45,440 yuan.

[0021] Trial molding and testing costs = trial molding material consumption + trial molding equipment energy consumption + testing fees + scrap disposal costs; where trial molding equipment energy consumption = 2 × 50 × 2 (trial molding time) = 200 yuan, scrap disposal costs = 500 yuan; final trial molding and testing costs = 2 × 1500 + 200 + 3000 + 500 = 6700 yuan.

[0022] Operation and maintenance costs and hidden costs = maintenance costs + fault repair costs + management costs + environmental compliance costs; where maintenance costs = (100,000 ÷ 1,000) × 800 = 80,000 yuan, fault repair costs = 100,000 × 0.5% × 1,000 (single fault repair cost) = 50,000 yuan, management costs = (design costs + material costs + processing costs + trial molding and testing costs) × 5% = 8,107.44 yuan, and environmental compliance costs are quantified as 113,040 yuan through the Softcom "Dual Carbon" control module; the final operation and maintenance costs and hidden costs = 80,000 + 50,000 + 8,107.44 + 113,040 = 251,147.44 yuan.

[0023] Step 3: Construct the iSoftStone AI intelligent cost prediction model to achieve accurate prediction; Relying on the iSoftStone AI algorithm platform, the model is trained using the standardized dataset from Step 1, and the costs of each dimension decomposed in Step 2 are predicted separately. The total predicted value of mold manufacturing cost is calculated by weighted summation. Combined with the simulation and deduction capabilities of the iSoftStone iSSMeta platform, the prediction error is corrected to ensure that the prediction error is controlled within 5%. The process of training a model using a standardized dataset includes: dividing the standardized dataset into a training set, a validation set, and a test set; training the model using the training set; adjusting the model parameters using the validation set; calculating the prediction error using the test set; and correcting the error using digital twin simulation capabilities. The weights of each cost dimension are obtained through training on historical data from the Softcom Industrial Big Data Platform. Specifically: Data training: The standardized dataset is divided into a training set (70%), a validation set (20%), and a test set (10%). The AI ​​prediction model is trained using the training set, and the model's learning rate, number of iterations, and other parameters are adjusted using the validation set to optimize the model's prediction accuracy.

[0024] Dimensional Prediction: Using a trained AI model, the cost of five dimensions is predicted separately, with the following predicted values: design cost 16,500 yuan, material cost 93,500 yuan, processing cost 45,200 yuan, trial molding and testing cost 6,600 yuan, and operation and maintenance and hidden cost 250,800 yuan.

[0025] Comprehensive forecast: The cost weights for each dimension were obtained by training on historical data from the Softcom Industrial Big Data Platform, and are as follows: design cost 3.8%, material cost 21.9%, processing cost 10.6%, trial molding and testing cost 1.5%, and operation and maintenance and hidden costs 62.2%. The total predicted cost is approximately RMB 427,000, calculated as follows: 16,500 × 3.8% + 93,500 × 21.9% + 45,200 × 10.6% + 6,600 × 1.5% + 250,800 × 62.2%.

[0026] Error correction: The test set data is input into the model to calculate the prediction error. Combined with the simulation and extrapolation capabilities of the iSSMeta platform, scenarios such as raw material price fluctuations and processing technology adjustments are simulated to dynamically correct the prediction model. The final prediction error is 0.9%, which is controlled within 5%.

[0027] Step 4: Build a dynamic iterative optimization mechanism to improve predictive adaptability; Relying on the real-time data update capability of the iSoftStone Industrial Cloud Platform, collect various dynamic variables in real time, update the basic dataset, and combine the error data between the actual manufacturing cost and the predicted cost of the mold to dynamically iterate and optimize the AI ​​prediction model. At the same time, realize multi-terminal collaborative feedback through the iSoftStone Collaborative Management Platform to adapt to industry technology iteration and market changes. The dynamic iterative optimization mechanism includes four stages: real-time data updates, model iterative optimization, scenario adaptive adjustment, and multi-terminal collaborative feedback. Real-time data updates synchronously collect information on changes in raw material prices, equipment status, and industry benchmark data. Model iterative optimization uses the error data between actual and predicted costs to adjust model parameters and cost weights. Scenario adaptive adjustment uses a digital twin platform to simulate new manufacturing scenarios and optimize model parameters. Multi-terminal collaborative feedback enables feedback and interaction among personnel at each stage through the Softcom collaborative management platform.

[0028] Specifically, real-time data updates: Softcom's AIoT module and industrial big data platform collect data in real time that the unit price of H13 hot work die steel has dropped to 18,400 yuan / ton and the processing time has been shortened to 115 hours, and update the basic dataset for cost prediction simultaneously.

[0029] Model Iteration and Optimization: After updating the data, the model re-predicts the total manufacturing cost to 425,100 yuan, with the prediction error corrected to 0.6%. After the mold is manufactured, the actual manufacturing cost is 424,800 yuan. The actual cost is compared with the predicted cost, and the calculation error is 0.07%. This error data is used to adjust the AI ​​prediction model parameters and the cost weights of each dimension.

[0030] Adaptive scenario adjustment: Through the iSSMeta digital twin platform of iSoftStone, the scenario of lightweight improvement of integrated die-casting mold is simulated, the model parameters are optimized, and the model is made suitable for new manufacturing requirements.

[0031] Multi-terminal collaborative feedback: Through the Softcom collaborative management platform, personnel from various stages such as design, procurement, processing, and operation and maintenance can provide feedback on abnormal cost data in actual production, further optimizing the accuracy of model predictions. Step 5: Output cost prediction results and optimization suggestions; Through the Softcom visualization module, output the cost prediction value, cost composition and error range, and combine with Softcom's digital optimization solution to provide targeted cost optimization suggestions in terms of materials, processes, design and operation and maintenance.

[0032] Multi-dimensional cost optimization recommendations include material optimization, process optimization, design optimization, and operation and maintenance optimization. Among them, material optimization recommends the most cost-effective material alternatives and procurement solutions, process optimization recommends the optimal processing technology route, design optimization recommends solutions to simplify mold structure and reduce design complexity, and operation and maintenance optimization recommends reasonable maintenance cycles and maintenance plans.

[0033] Specifically, material optimization is recommended: the price of H13 hot work die steel is currently declining, so it is suggested to increase the purchase volume to 8 tons, which can reduce the unit price of materials to 18,200 yuan / ton and reduce material costs by about 3,000 yuan; at the same time, it is recommended to use new energy-saving die steel to replace part of H13 steel to further reduce energy consumption and environmental protection costs.

[0034] Process optimization: By using digital twin simulation, the EDM machining process can be replaced with a high-speed milling process, which can shorten the machining time to 110 hours and reduce the machining cost by about 2,200 yuan.

[0035] Design optimization: Simplifying non-critical structures of the mold reduces the design complexity score to 8.0, which can reduce design costs by about 600 yuan and shorten the design cycle by 2 days.

[0036] Operation and maintenance optimization: Adjusting the mold maintenance cycle to 1200 mold cycles can reduce maintenance costs by approximately 16,000 yuan; at the same time, through the predictive maintenance module of Softcom, equipment failures can be warned in advance, reducing failure repair costs.

[0037] The cost prediction method relies on iSSMeta digital simulation and deduction platform, industrial big data platform, AI algorithm platform, AIoT module, MES production and manufacturing execution platform, collaborative management platform, and "dual carbon" control module of iSoftStone, and can be seamlessly connected with the existing production system of auto parts manufacturing enterprises.

[0038] This technical solution targets the complex integrated die-casting molds for new energy vehicles. The method employed in this invention achieves precise prediction of manufacturing costs with an error margin of only 0.6%, far lower than the existing technology's error rate of over 15%, providing accurate data support for enterprise quotations. Through a dynamic iterative optimization mechanism, it adapts in real-time to variables such as raw material price fluctuations and process adjustments, ensuring the timeliness and accuracy of the prediction results. Targeted cost optimization suggestions help enterprises reduce manufacturing costs by approximately 23,800 yuan, a reduction of 5.6%. Simultaneously, relying on the Softcom digital platform, it achieves visualized management of the mold's entire lifecycle cost, shortening the cost prediction cycle (from the traditional 3-5 days to within 1 day), improving enterprise order response efficiency, and solving the industry pain point of accurately breaking down and predicting the costs of complex die-casting molds.

[0039] Example 2 This embodiment focuses on the die-casting mold for a traditional gasoline-powered automotive engine block (material: gray cast iron, displacement: 2.0L). The mold structure is relatively conventional, containing four cavities, with a dimensional accuracy requirement of ≤±0.05mm and a service life of ≥80,000 cycles. Enterprises need to optimize procurement and production plans through cost forecasting to control production costs. This embodiment relies on the Softcom Digital Platform, with data sourced from actual enterprise production data and Softcom's industry benchmark database.

[0040] Specific implementation steps Step 1: Construct a digital twin of the soft-pass die-casting mold and collect basic data throughout its entire lifecycle. Based on the iSSMeta digital simulation platform of iSoftStone, a 1:1 digital twin of the engine block die-casting mold was constructed. Basic data from all stages of the entire life cycle were collected, processed by iSoftStone's data cleaning and standardization module, and stored on the iSoftStone industrial cloud platform. The specific data is as follows: Design data: The mold 3D model was imported through the Softcom CAD collaborative design module. The improved BREPNet architecture and MoE hybrid expert model were used to automatically identify key structures such as mold cavities and gates. The quantitative design complexity score was 6.0. The design cycle was 10 days, the designer's salary was 280 yuan / day, and there was no design iteration.

[0041] Material data: Through the Softcom Industrial Big Data Platform, we connected with Cr12MoV mold steel suppliers and collected real-time material unit price data. The unit price was 15,800 yuan / ton, the purchase batch was 3 tons, the transportation cost was 500 yuan, and the loss rate was 2.5%. The mold material usage was automatically calculated to be 2.7 tons through the digital twin 3D model.

[0042] Processing data: Through the connection of the processing equipment via the Softcom AIoT module, the processing time is estimated to be 80 hours, the equipment depreciation is 40 yuan per hour, the energy consumption is 12 kWh per hour (electricity price is 0.8 yuan / kWh), and the tool wear is estimated to be 1500 yuan; the processing accuracy data is collected synchronously through the Softcom MES platform.

[0043] Trial molding and maintenance data: The trial molding process is simulated by a digital twin. The number of trial molding tests is 1, the material consumption of trial molding is 1200 yuan, and the testing cost is 2000 yuan. Through the integration with the Softcom maintenance system, the mold maintenance cycle is collected as 1500 mold cycles, the cost of each maintenance is 600 yuan, and the failure probability is 0.3%.

[0044] Industry benchmark data: By aggregating historical cost data, industry average loss rate (2.3%), and average maintenance cost (580 yuan / time) of similar engine cylinder block die-casting molds through the Softcom Industrial Big Data Platform.

[0045] Step 2: Construct a multi-dimensional cost decomposition model for Softcom Based on iSoftStone's industrial big data analytics capabilities, the manufacturing cost of this mold is broken down into five core dimensions and dynamically calculated as follows: Design cost = 6.0 × 1000 (design cost per unit of complexity, adjusted based on industry benchmark) + (10 × 280) = 6000 + 2800 = 8800 yuan.

[0046] Material cost = (2.7 × (1 + 2.5%) × 15800) + 500 + (2.7 × 15800 × 0.5%) = 43285.5 + 213.3 = 43498.8 yuan.

[0047] Processing cost = 80 × 280 (processing personnel salary) + 80 × 40 + 80 × 12 × 0.8 + 1500 = 22400 + 3200 + 768 + 1500 = 27868 yuan.

[0048] Trial molding and testing costs = 1200 + 80 (energy consumption of trial molding equipment) + 2000 + 300 (cost of scrap disposal) = 3580 yuan.

[0049] Operation and maintenance costs and hidden costs = (80000÷1500)×600+(80000×0.3%×800)+(8800+43498.8+27868+3580)×4%+78000 (environmental compliance costs)=32000+19200+3349.87+78000=132549.87 yuan.

[0050] Step 3: Build a Softcom AI intelligent cost prediction model to achieve accurate predictions. Based on the iSoftStone AI algorithm platform, integrating LSTM neural network and random forest algorithm, and training the model using standardized dataset, the predicted cost values ​​for each dimension are as follows: design cost 8700 yuan, material cost 43600 yuan, processing cost 27700 yuan, trial molding and testing cost 3600 yuan, and operation and maintenance and hidden cost 132300 yuan. The cost weights for each dimension are 4.2% for design cost, 21.1% for material cost, 13.4% for processing cost, 1.7% for trial molding and testing cost, and 59.6% for operation and maintenance and hidden cost. The total predicted cost is approximately 206000 yuan (8700×4.2%+43600×21.1%+27700×13.4%+3600×1.7%+132300×59.6%). After error correction, the prediction error is 0.7%, which is controlled within 5%.

[0051] Step 4: Construct a dynamic iterative optimization mechanism to improve predictive adaptability The iSoftStone Industrial Cloud Platform collected real-time data showing that the unit price of Cr12MoV mold steel had risen to 16,000 yuan / ton, and the energy consumption of processing equipment had decreased by 10%. After the model updated the data in real time, it re-predicted the total manufacturing cost to be 207,200 yuan, with the prediction error corrected to 0.5%. After the mold was manufactured, the actual manufacturing cost was 207,000 yuan. Using the error data between the actual and predicted costs, the AI ​​model parameters and cost weights were iteratively optimized. At the same time, through the iSoftStone collaborative management platform, feedback from each link was collected to further improve the model's adaptability.

[0052] Step 5: Output cost prediction results and optimization suggestions Through the Softcom visualization module, the total predicted cost of 207,200 yuan is output, along with the cost composition and error range for each dimension. Combined with Softcom's digital optimization solution, optimization suggestions are provided: In terms of material optimization, it is recommended to sign long-term supply agreements with suppliers to lock in material prices and reduce the risk of price fluctuations; in terms of process optimization, optimize the processing path, shorten the processing time to 75 hours, and reduce processing costs by approximately 1,400 yuan; in terms of operation and maintenance optimization, optimize the maintenance process, reduce maintenance costs to 550 yuan per instance, and reduce operation and maintenance costs by approximately 2,600 yuan.

[0053] This technical solution targets die-casting molds for traditional gasoline-powered automotive engine blocks. The method employed in this invention achieves accurate prediction of manufacturing costs with a prediction error of only 0.5%, overcoming the problems of high subjectivity and low accuracy associated with traditional manual prediction. The dynamic iterative optimization mechanism responds in real-time to variables such as raw material price fluctuations and equipment energy consumption changes, ensuring the accuracy and timeliness of the prediction results. Through targeted cost optimization suggestions, it helps enterprises reduce manufacturing costs by approximately 4,000 yuan, a reduction of 1.9%. Simultaneously, relying on the Softcom digital platform, it achieves full-process control of mold costs, reducing issues such as mold trial wear and material waste, improving enterprise production efficiency, and adapting to the mass production needs of traditional automotive parts die-casting molds. It possesses significant practicality and promotional value.

[0054] Example 3 This embodiment focuses on the die-casting mold for a new energy vehicle battery casing (material: aluminum alloy, thickness: 1.2mm). This mold requires lightweight design and high precision, incorporating multiple core-pulling structures, with a dimensional accuracy of ≤±0.03mm and a service life of ≥120,000 cycles. Enterprises need to balance cost and product quality through accurate cost forecasting to enhance market competitiveness. This embodiment relies on the Softcom Digital Platform, with data sourced from actual production scenarios and industry benchmark data.

[0055] Specific implementation steps The implementation process for steps one through five is the same as in Examples 1 and 2. The core differences lie in the basic data, cost calculation parameters, and optimization suggestions, as detailed below: Basic data: Design complexity score 7.5, design cycle 12 days, design iteration 1 time; material used is ADC12 aluminum alloy mold steel, unit price 17,200 yuan / ton, purchase batch 4 tons, loss rate 2.8%; processing time 100 hours, equipment depreciation 45 yuan / hour, energy consumption 14 kWh / hour; number of trial moldings 2 times, trial molding material loss 1,300 yuan / time; maintenance cycle 800 mold cycles, maintenance cost 700 yuan per cycle.

[0056] Multi-dimensional cost breakdown: Design cost = 7.5 × 1100 + (12 × 290) + 1800 = 12380 yuan; Material cost = (3.8 × (1 + 2.8%) × 17200) + 600 + 295.68 = 68234.88 yuan; Processing cost = 100 × 290 + 100 × 45 + 100 × 14 × 0.8 + 1800 = 37420 yuan; Trial molding and testing cost = 2 × 1300+180+2500+400=5680 yuan; Operation and maintenance and hidden costs = (120000÷800)×700+(120000×0.4%×900)+(12380+68234.88+37420+5680)×4.5%+95000=105000+43200+5566.57+95000=248766.57 yuan.

[0057] AI prediction and error correction: Total prediction cost = 12380×3.5% + 68234.88×19.2% + 37420×10.6% + 5680×1.6% + 248766.57×65.1% ≈ 352000 yuan, prediction error 0.8%.

[0058] Optimization suggestions: Material optimization: High-strength, lightweight aluminum alloy is recommended to reduce mold weight and material usage; Process optimization: Precision die casting process is recommended to reduce the number of trial moldings; Design optimization: Simplify the core-pulling structure to reduce design complexity; Operation and maintenance optimization: Intelligent maintenance system is recommended to extend mold life.

[0059] This technical solution targets die-casting molds for new energy vehicle battery casings. Through this invention, it achieves accurate cost prediction for lightweight, high-precision molds, with prediction errors controlled within 1%, meeting enterprises' needs for cost control of high-end precision molds. A dynamic iterative optimization mechanism adapts to the development trend of lightweight and precision in new energy vehicle parts, ensuring the long-term applicability of the prediction model. Targeted cost optimization suggestions help enterprises reduce manufacturing costs by approximately 18,000 yuan, a reduction of 5.1%, while simultaneously improving mold lifespan and product quality. Relying on the Softcom digital platform, it achieves collaborative management of the entire mold design, processing, and maintenance process, shortening the production cycle, enhancing the enterprise's core competitiveness, and providing a feasible solution for cost control of die-casting molds for new energy vehicle parts.

[0060] The above description of the embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the manufacturing cost of die-casting molds for automotive parts, characterized in that, This includes iSoftStone's industrial metaverse, digital twin, and industrial big data technologies. The specific steps are as follows: Step 1: Construct a digital twin of the iSSMeta die-casting mold and collect basic data throughout the entire lifecycle; Relying on the iSSMeta digital simulation and deduction platform of iSoftStone, construct a 1:1 digital twin of the die-casting mold of automotive parts, collect basic data of each stage of the entire lifecycle, and store the data on the iSoftStone industrial cloud platform after iSoftStone data cleaning and standardization. Step 2: Construct a multi-dimensional cost decomposition model based on iSoftStone's industrial big data capabilities; break down mold costs into five core dimensions, use dynamic calculation methods, and combine industry benchmark data to achieve refined decomposition. Step 3: Construct the iSoftStone AI intelligent cost prediction model to achieve accurate predictions; relying on the iSoftStone AI algorithm platform, integrating LSTM and random forest algorithms, training the model using a standardized dataset, predicting costs in each dimension and summing them with weights, and using the iSSMeta platform to correct errors, ensuring that the error is ≤5%; Step 4: Build a dynamic iterative optimization mechanism to improve predictive adaptability; rely on real-time data updates from iSoftStone Industrial Cloud, combine actual and predicted cost errors to iteratively optimize the AI ​​model, and achieve multi-terminal feedback through the iSoftStone collaborative management platform to adapt to industry changes; Step 5: Output cost forecast results and optimization suggestions; Output cost forecast results and their composition through the iSoftStone visualization module, and provide multi-dimensional cost optimization suggestions in conjunction with iSoftStone's digital solutions.

2. The method for predicting the manufacturing cost of automotive parts die-casting molds according to claim 1, characterized in that, The aforementioned basic data includes design data, material data, processing data, trial molding and maintenance data, and industry benchmark data; The design data is imported through the iSoftStone CAD collaborative design module, which uses the improved BREPNet architecture and MoE hybrid expert model to automatically identify key mold structures and quantify design complexity scores; material data is collected in real time from upstream suppliers through the iSoftStone Industrial Big Data Platform; processing data is collected in real time from processing equipment through the iSoftStone AIoT module; trial molding and operation and maintenance data are collected through digital twin simulation and operation and maintenance system integration; and industry benchmark data is generated by aggregating historical industry data through the iSoftStone Industrial Big Data Platform.

3. The method for predicting the manufacturing cost of automotive parts die-casting molds according to claim 1, characterized in that, The dynamic calculation method is specifically as follows: Design cost = Design complexity score × Design cost per unit complexity + Design iteration cost. The design cost per unit complexity is based on data from the Softcom industry benchmark database and is dynamically adjusted in conjunction with the design cycle and designers' salaries. Material cost = (mold material usage × (1 + loss rate) × real-time material unit price) + transportation cost + warehousing cost. Mold material usage is automatically calculated through a digital twin 3D model, and the loss rate is dynamically optimized based on historical data from the iSoftStone Industrial Big Data Platform. Processing cost = labor cost + equipment depreciation cost + energy cost + tooling cost. Labor cost is dynamically calculated by simulating the processing process through a digital twin. Equipment depreciation and energy cost are calculated based on real-time data collected by the Softcom AIoT module. Trial molding and testing costs = trial molding material loss + trial molding equipment energy consumption + testing fees + scrap disposal costs. The number of trial molding tests and losses can be predicted by simulating the trial molding process using a digital twin. Operation and maintenance costs and hidden costs = maintenance costs + fault repair costs + management costs + environmental compliance costs. Maintenance costs are based on the mold's lifespan and maintenance cycle prediction, while environmental compliance costs are quantified through the Softcom dual-carbon control module.

4. The method for predicting the manufacturing cost of automotive parts die-casting molds according to claim 1, characterized in that, The standardized dataset training model process includes: dividing the standardized dataset into a training set, a validation set, and a test set; training the model using the training set; adjusting the model parameters using the validation set; calculating the prediction error using the test set; and correcting the error using digital twin simulation capabilities. The weights of each cost dimension are obtained through training with historical data from the Softcom Industrial Big Data Platform.

5. The method for predicting the manufacturing cost of automotive parts die-casting molds according to claim 1, characterized in that, The dynamic iterative optimization mechanism includes four stages: real-time data update, model iterative optimization, scenario adaptive adjustment, and multi-terminal collaborative feedback; real-time data update synchronously collects information on changes in raw material prices, equipment status, and industry benchmark data. Model iterative optimization uses the error data between actual and predicted costs to adjust model parameters and cost weights; Scenario adaptive adjustment simulates new manufacturing scenarios through a digital twin platform to optimize model parameters; multi-terminal collaborative feedback enables feedback and interaction among personnel at all stages through the Softcom collaborative management platform.

6. The method for predicting the manufacturing cost of automotive parts die-casting molds according to claim 1, characterized in that, The multi-dimensional cost optimization suggestions include material optimization, process optimization, design optimization, and operation and maintenance optimization. Among them, material optimization recommends the most cost-effective material alternatives and procurement solutions, process optimization recommends the optimal processing technology route, design optimization recommends solutions to simplify mold structure and reduce design complexity, and operation and maintenance optimization recommends reasonable maintenance cycles and maintenance plans.

7. A method for predicting the manufacturing cost of automotive parts die-casting molds according to any one of claims 1-6, characterized in that, The cost prediction method relies on iSSMeta digital simulation and deduction platform, industrial big data platform, AI algorithm platform, AIoT module, MES production and manufacturing execution platform, collaborative management platform, and dual carbon control module of iSoftStone, and can be seamlessly connected with the existing production system of automotive parts manufacturing enterprises.

Images