Outsourcing processing cost calculation method based on labor-material ratio and dynamic management system

Abstract

The invention relates to the technical field of foreign unit entrusted processing management, and discloses a foreign unit entrusted processing cost calculation method based on a labor-to-material ratio, and the method comprises the following steps: 1, building a cooperative unit directory and a contact information manual; 2, constructing a partner technology strength assessment index system and a risk assessment model; step 3, formulating a classification standard of the machined part; 4, determining a labor-material ratio coefficient, and calculating the processing cost by adopting the coefficient according to a classification result; 5, signing a one-year standardized contract; and 6, constructing a machined part quality evaluation model. According to the outsourcing processing cost calculation method based on the labor-material ratio and the dynamic management system, a partner management module establishes a classification ranking directory and a quantitative evaluation model in advance to lock a preferential partner, and a processing part classification module rapidly outputs a classification result through automatic detection and a classification model. The contract module provides a standardized template to realize rapid signing, and all the modules cooperate to omit dispersed drawing audit and bid invitation links.

Description

Technical Field

[0001] This invention relates to the field of outsourced processing management technology, specifically a method for calculating outsourced processing costs based on the material-to-labor ratio and a dynamic management system. Background Technology

[0002] In recent decades, my country's thermal power industry has seen rapid technological development, with new technologies and equipment constantly emerging, placing higher demands on equipment management and technology management. The scope of processing services for production equipment in thermal power enterprises is expanding, and their reliance on outsourced processing is increasing year by year. Outsourcing non-core component customization, equipment repair parts processing, and special tooling and fixture processing to specialized external units has become an important way to supplement production capacity, optimize resource allocation, and reduce investment in fixed assets such as factory buildings and equipment, as well as the operational pressure on core businesses.

[0003] However, the existing management of outsourced processing faces three major technical problems:

[0004] 1. The processing cycle is long. The traditional process involves multiple scattered steps such as drawing review, bidding, processing, and transportation. It cannot respond quickly to urgent production needs, with an average cycle of more than 7 days, which seriously affects the continuity of production and may even lead to production line shutdown.

[0005] 2. Lack of transparency in pricing and a lack of scientific and unified pricing standards make it easy for partners to "raise prices on the spot" during emergency bidding. There is no uniform standard for quotations from different partners, making it difficult for enterprises to effectively control processing costs and adding extra operating burdens.

[0006] 3. Insufficient quality and risk control. The selection of partners relies on subjective human judgment and lacks a dynamic and quantitative evaluation mechanism. Some partners have issues such as qualification fraud and unqualified personnel, resulting in inconsistent quality of processed parts, lack of after-sales support, and easy to cause production safety accidents and economic losses.

[0007] Therefore, a method for calculating outsourced processing costs based on the material-to-labor ratio and a dynamic management system are proposed. Summary of the Invention

[0008] (a) Technical problems to be solved

[0009] To address the shortcomings of existing technologies, this invention provides a method and dynamic management system for calculating outsourced processing costs based on the ratio of labor and materials. This system features an outsourced processing management scheme that achieves advantages such as improved processing efficiency, controllable costs, and reduced risks through "standardized classification, scientific pricing, quantitative evaluation, and full-cycle control." It solves the problems of long processing cycles, opaque costs, and insufficient quality and risk control in outsourced processing management technologies.

[0010] (II) Technical Solution

[0011] To achieve the aforementioned outsourced processing management scheme, which utilizes "standardized classification, scientific pricing, quantified evaluation, and full-cycle control" to improve processing efficiency, controllable costs, and reduced risks, this invention provides the following technical solution: A method for calculating outsourced processing costs based on the material-labor ratio, comprising the following steps:

[0012] Step 1: Establish a directory of partner organizations and a contact information booklet, categorized and ranked by profession;

[0013] Step 2: Construct a technical strength evaluation index system and risk assessment model for partners, conduct regular surveys on partners' technical strength, financial strength, and social reputation, and determine the priority partners through model scoring;

[0014] Step 3: Establish classification standards for processed parts, classifying them into large and medium-sized parts according to size, and into three levels: simple, easy, and difficult according to material hardness, rigidity, precision, tolerance, and shape;

[0015] Step 4: Determine the material-labor ratio coefficient, and calculate the processing cost using the coefficient based on the classification results. The formula is: ;

[0016] The formula for processing costs with supplied materials is: ;

[0017] in This is the gross weight of the material. This is the unit price of the material. The difficulty ratio and , Profit coefficient and , Tax coefficient and , For shipping costs;

[0018] Step 5: Sign a one-year standardized contract, clearly defining rights and obligations, price agreement, and after-sales service methods;

[0019] Step Six: Construct a quality evaluation model for processed parts, establish a quality responsibility mechanism for processed parts, reserve a quality guarantee deposit, and implement standardized management of the entire process through a division of labor involving dedicated personnel for communication, technical standard formulation, and business plan supervision.

[0020] Preferably, the process for constructing the partner's technical strength evaluation index system in step two is as follows:

[0021] Step 1: Select three core indicators: operator skill level, number of processing equipment, and establishment time;

[0022] Step 2: Use the analytic hierarchy process (AHP) to determine the weights of the indicators: operator skill level weight 0.4, number of processing equipment weight 0.3, and establishment time weight 0.3.

[0023] Step 3: Collect historical data from no fewer than 50 partners. This sample size meets the minimum sample requirement of the random forest model after statistical testing. Divide the data into an 8:2 training set and a test set. Train the model using the random forest algorithm and iterate 500 times until the loss value is below 0.05.

[0024] Step 4: Ensure the model accuracy is no less than 85% through 5-fold cross-validation;

[0025] Step 5: Input partner data in real time, and the model outputs a technical strength score: 5 points for technician, 3 points for senior worker, 2 points for intermediate worker, 1 point for junior worker, 5 points for more than 5 pieces of equipment, 1 point for less than 3 pieces of equipment, no points for no equipment, 5 points for establishment time of more than 5 years, 2 points for establishment time of less than 3 years, and no points for newly established companies.

[0026] Preferably, the process for constructing and executing the workpiece classification standard in step three is as follows:

[0027] Step 1: The size classification standard is determined based on the finished size of the processed parts. Plates with a length > 50mm are large parts, and those ≤ 50mm are small to medium parts. Rods with a diameter > 45mm are large parts, and those ≤ 45mm are small to medium parts. Pipes with an outer diameter > 50mm are large parts, and those ≤ 50mm are small to medium parts. The basis for using the corresponding classification standard for different types of parts is the difference in processing technology and industry processing practices.

[0028] Step 2: The difficulty level standards are as follows: material hardness HRC0-HRC20 is easy, HRC20-HRC40 is easy, and ≥HRC40 is difficult; length <500mm and outer diameter >150mm is easy, length <1000mm and 50mm < outer diameter <100mm is easy, and length >1000mm and 10mm < outer diameter <50mm is difficult; Ra100, Ra50, Ra25 are easy, Ra12.5, Ra6.3, Ra3.2 are easy, and Ra1.6, Ra0.8, Ra0.4 are difficult; free tolerance is easy, ±0.05 is easy, and ±0.01 is difficult; semi-enclosed cavity box-type parts are classified as difficult.

[0029] Step 3: When a workpiece simultaneously meets multiple difficulty level characteristics, the most difficult level shall be used to determine the difficulty level.

[0030] Step 4: Identify the dimensions of the finished workpiece using machine vision, test the material hardness using a hardness tester, and input the results into the classification model to automatically output the classification results;

[0031] Step 5: Calculate the gross weight of the bar stock according to the specification of the largest dimension above the maximum dimension in the processing drawing. The difference between the classification standard and the material calculation rule is based on the industry standard for waste loss in bar stock processing.

[0032] Preferably, the process for determining the material-to-labor ratio coefficient and operating the pricing model in step four is as follows:

[0033] Step 1: Collect 100 sets of historical data on machined parts, and fit the difficulty ratio using a linear regression algorithm. Based on the verification and determination of coefficient values ​​by 5 industry experts, the difficulty of large items is simplified. Easy Difficult Small and medium-sized items are simple in difficulty Easy Difficult ;

[0034] Step 2: Collect no fewer than 30 sets of new processing data each quarter, and update the coefficients and profit coefficients using the gradient descent method. Initially 1.1, later adjusted to 1.08, tax coefficient. The initial value was 1.17, which was later adjusted to 1.13. The adjustment was based on national tax policies and changes in the industry's average profit margin.

[0035] Step 3: Enter the gross weight of materials Material unit price Based on the classification results, the model automatically calculates the processing costs.

[0036] Step 4: "≥L800(φ800)" means that the length of the processed part is ≥800mm or the diameter is ≥800mm. At this time, the material cost will increase by 3% of the scrap ratio, based on the actual measured data of the scrap loss rate of large parts processing.

[0037] Step 5: Shipping costs per batch The amount is determined based on the average transportation cost of processed parts within the region.

[0038] Preferably, the process for constructing and operating the workpiece quality evaluation model in step six is ​​as follows:

[0039] Step 1: Select three indicators: service life, machining accuracy, and failure rate, with weights of 0.4, 0.3, and 0.3 respectively, and determine them using the analytic hierarchy process (AHP).

[0040] Step 2: Collect operational data from 200 processed parts, train the model using the fuzzy comprehensive evaluation method, and set five levels from "five-star rating" to "product unusable", with each level corresponding to a specific indicator threshold;

[0041] Step 3: Collect real-time data on the operation of the processed parts after installation, and output evaluation results from the model. If the parts are damaged before the agreed operating cycle is reached, the partner will replace them free of charge. If the damage exceeds three times, the deposit will be withheld and the cooperation will be terminated.

[0042] Preferably, the process for constructing and applying the partner risk assessment model in step two is as follows:

[0043] Step 1: Construct a four-level evaluation index system, including project management risk, contractor management risk, project performance risk, and project quality risk. Each primary index contains 3-4 secondary indicators.

[0044] Step 2: Invite 10 industry experts to score, use the analytic hierarchy process to determine the weights of each level of indicators, combine with the fuzzy comprehensive evaluation model for training, and iteratively optimize until the consistency test CR < 0.1;

[0045] Step 3: The model outputs risk levels, including very high risk, high risk, medium risk, and low risk. Cooperation is possible for medium risk and below, and process monitoring should be strengthened. Cooperation is excluded for very high risk. The risk level classification is based on the risk occurrence probability and impact matrix.

[0046] A dynamic management system for outsourced processing costs based on the material-to-labor ratio, comprising:

[0047] Partner Management Module: Stores partner directory and evaluation data, and deploys partner technical strength evaluation index system and risk assessment model;

[0048] Processed Parts Classification Module: Deploys the processing parts classification standards and models, receives processing parts parameters and outputs classification results, which are then output to the pricing module as input parameters.

[0049] Pricing module: Deploys the material cost ratio and pricing model, inputs the gross weight of materials, unit price of materials and classification results, and automatically calculates processing costs;

[0050] Contract module: Stores standardized contract templates, supports contract generation and signing process management, and synchronizes contract data to the full lifecycle management module;

[0051] Full-cycle management module: Deploys a processing part quality evaluation model, records processing part operation data, manages quality assurance deposits, and synchronizes data to the query platform sub-module;

[0052] Data interaction module: Connects to external material price data, supports various data input and export, and includes a contract information extraction unit.

[0053] Preferably, the contract information extraction unit process of the data interaction module is as follows:

[0054] Step 1: Build a regular expression matching library containing keywords for manufacturing name, contractor, and signing date. The keyword library was determined by analyzing 100 samples of outsourcing contracts.

[0055] Step 2: Train the fuzzy inference filtering model. The truth value calculation formula is:

[0056] in The weight of the premise evidence is set, with a value range of [0.2, 0.5], based on the importance of the evidence.

[0057] Step 3: Input the outsourcing contract, first extract candidate information through regular expression matching, then use fuzzy reasoning to filter the model to calculate the truth value, and output the key information with the highest truth value.

[0058] Preferably, the query platform sub-module displays information such as processing part classification, partner unit ranking, processing cycle, and price prediction. The data is updated hourly to ensure information timeliness. The data comes from synchronized data from various functional modules and is displayed after consistency verification.

[0059] (III) Beneficial Effects

[0060] Compared with the prior art, the present invention provides a method and dynamic management system for calculating outsourced processing costs based on the material-labor ratio, which has the following beneficial effects:

[0061] 1. This outsourcing processing cost calculation method and dynamic management system based on the material-labor ratio allows the partner management module to pre-establish a categorized and ranked directory and a quantitative evaluation model to lock in preferred partners. The processing parts classification module quickly outputs classification results through automated detection and classification models. The contract module provides standardized templates to enable rapid signing. The collaboration of these modules eliminates the need for scattered drawing review and bidding processes, meets the requirements for rapid response, and shortens the processing cycle.

[0062] 2. This outsourcing processing cost calculation method and dynamic management system based on the material-labor ratio has a processing parts classification module that classifies parts by size and difficulty level according to a unified standard, and a pricing module that automatically calculates costs based on the material-labor ratio coefficient, scientific formulas, and dynamically updated parameters, forming a unified pricing basis. This avoids arbitrary pricing by partners, allows companies to clearly control costs, and reduces their operating burden.

[0063] 3. The outsourcing processing cost calculation method and dynamic management system based on the material-labor ratio, the partner management module quantitatively screens partners through technical strength and risk assessment models, the full-cycle control module deploys a quality evaluation model to collect data in real time to evaluate quality, and with the addition of quality assurance deposits and responsibility mechanisms, the drawbacks of subjective screening are eliminated, ensuring the quality of processed parts and after-sales service, and reducing the risk of safety accidents. Attached Figure Description

[0064] Figure 1 This is a flowchart of the outsourcing processing cost calculation method based on the material-to-labor ratio of the present invention;

[0065] Figure 2 This is a diagram illustrating the architecture of the dynamic management system for outsourced processing costs based on the material-to-labor ratio of the present invention.

[0066] Figure 3 This is a data interaction flow diagram of the present invention. Detailed Implementation

[0067] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments and 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.

[0068] Please see Figure 1-3 A method for calculating outsourced processing costs based on the material-to-labor ratio includes the following steps:

[0069] Step 1: Establish a directory of partner organizations and a contact information booklet, categorized and ranked by profession;

[0070] Step 2: Construct a technical strength evaluation index system and risk assessment model for partners, conduct regular surveys on partners' technical strength, financial strength, and social reputation, and determine the priority partners through model scoring;

[0071] Step 3: Establish classification standards for processed parts, classifying them into large and medium-sized parts according to size, and into three levels: simple, easy, and difficult according to material hardness, rigidity, precision, tolerance, and shape;

[0072] Step 4: Determine the material-labor ratio coefficient, and calculate the processing cost using the coefficient based on the classification results. The formula is: ;

[0073] The formula for processing costs with supplied materials is: ;

[0074] in This is the gross weight of the material. This is the unit price of the material. The difficulty ratio and , Profit coefficient and , Tax coefficient and , For shipping costs;

[0075] Step 5: Sign a one-year standardized contract, clearly defining rights and obligations, price agreement, and after-sales service methods;

[0076] Step Six: Construct a quality evaluation model for processed parts, establish a quality responsibility mechanism for processed parts, reserve a quality guarantee deposit, and implement standardized management of the entire process through a division of labor involving dedicated personnel for communication, technical standard formulation, and business plan supervision.

[0077] A dynamic management system for outsourced processing costs based on the material-to-labor ratio, comprising:

[0078] Partner Management Module: Stores partner directory and evaluation data, and deploys partner technical strength evaluation index system and risk assessment model;

[0079] Processed Parts Classification Module: Deploys the processing parts classification standards and models, receives processing parts parameters and outputs classification results, which are then output to the pricing module as input parameters.

[0080] Pricing module: Deploys the material cost ratio and pricing model, inputs the gross weight of materials, unit price of materials and classification results, and automatically calculates processing costs;

[0081] Contract module: Stores standardized contract templates, supports contract generation and signing process management, and synchronizes contract data to the full lifecycle management module;

[0082] Full-cycle management module: Deploys a processing part quality evaluation model, records processing part operation data, manages quality assurance deposits, and synchronizes data to the query platform sub-module;

[0083] Data interaction module: Connects to external material price data, supports various data input and export, and includes a contract information extraction unit.

[0084] Example 1:

[0085] This embodiment details the specific implementation process and operational details of the technical features of the establishment of the list of cooperative units, the assessment of technical strength, and the risk assessment model.

[0086] First, establish a static information management system for partner organizations;

[0087] Through a three-month market survey, information on enterprises with processing capabilities in the region and industry was collected, covering core data such as production capacity, processing scale, and historical cooperation records, and a directory of cooperative units was established.

[0088] The list is divided into three categories according to processing specialties: custom parts processing, equipment repair and parts processing, and special tooling and fixture processing. Within each category, the list is ranked by processing level and cost standards. The ranking coefficient is calculated by weighting the processing accuracy compliance rate (0.4), historical cooperation satisfaction (0.3), and price reasonableness (0.3).

[0089] At the same time, a contact information manual for partners was compiled, distinguishing between emergency contacts, general contacts, and business contacts in the directory. Emergency contacts are responsible for responding to sudden processing needs 24 hours a day, while business contacts are specifically responsible for contract signing and payment settlement matters, reducing communication redundancy.

[0090] Second, we will construct an evaluation index system for the technical capabilities of our partners;

[0091] Three core indicators were selected: operator skill level, number of processing equipment, and establishment time. The weights were determined using the analytic hierarchy process (AHP), with operator skill level having a weight of 0.4, number of processing equipment having a weight of 0.3, and establishment time having a weight of 0.3.

[0092] Historical data from 60 partners were collected. This sample size was statistically verified to meet the minimum sample requirements for model training. The data was divided into a training set of 48 partners and a test set of 12 partners in an 8:2 ratio. After preprocessing the data by filling missing values ​​and removing outliers, the data was input into the random forest algorithm for model construction. The number of decision trees was set to 100, the maximum depth to 10, and the minimum number of samples for node splitting to 2. The mean squared error was used as the loss function, and the number of iterations was set to 500 until the model loss value dropped to 0.04.

[0093] The model was validated using 5-fold cross-validation. The dataset was randomly divided into 5 subsets, with 4 subsets used as the training set and 1 subset used as the test set, repeated 5 times. The average accuracy was then calculated, and the final model accuracy reached 88%. The specific model scoring criteria are as follows:

[0094] Among the operators, each technician receives 5 points, each senior worker receives 3 points, each intermediate worker receives 2 points, and each junior worker receives 1 point. The average of all operators' scores is then calculated.

[0095] 5 or more processing equipment: 5 points; 3 to 5 equipment: 3 points; less than 3 equipment: 1 point; no equipment: no points.

[0096] 5 points are awarded for establishments that have been established for more than 5 years, 3 points for establishments that have been established for 3 to 5 years, 2 points for establishments that have been established for less than 3 years, and no points are awarded for newly established companies. The final score for technical strength is the weighted sum of the scores of the three indicators, and the priority partners are determined according to the scores from high to low.

[0097] After deployment, the model is embedded in the information management system and receives data updates from partners in real time via API. It is retrained every six months to adapt to market changes.

[0098] Third, we will then build a risk assessment model for our partners;

[0099] A four-level evaluation index system is constructed. The first-level indexes include project management risk, contractor management risk, project performance risk, and project quality risk. Each first-level index contains 3 to 4 second-level indexes. Project management risk includes three second-level indexes: standardization of construction process, completeness of safety management mechanism, and cost control capability.

[0100] Ten industry experts, each with over eight years of experience, were invited to score the weights of each indicator level. Based on the scoring results, pairwise comparison matrices were constructed. The weight vectors were calculated using the analytic hierarchy process (AHP) and a consistency test was performed. The consistency test formula is as follows: ; ;

[0101] in To determine the largest eigenvalue of a matrix, To determine the dimension of a matrix, The average random consistency index is used; the weights are combined with the fuzzy comprehensive evaluation model for training, and after 100 iterations of optimization, the consistency is checked. The value is 0.08, indicating that the model has converged.

[0102] The model outputs four levels: extremely high risk, high risk, medium risk, and low risk. Partners with medium risk or below are included in the candidate pool and their process is monitored more closely. Partners with extremely high risk are directly excluded from the cooperation. The risk level classification is determined based on the risk occurrence probability and impact matrix. Partners that have had more than 3 quality incidents are judged as extremely high risk, and partners that have had 1 quality incident but have completed rectification are judged as medium risk.

[0103] After the model is deployed, it receives dynamic data from partners in real time and updates the risk level assessment results every quarter.

[0104] This embodiment establishes a standardized directory of partner organizations and a manual of their contact information. By combining multi-dimensional indicators and algorithm models to construct an evaluation system, it achieves precision and standardization in partner selection, reduces communication redundancy and subjective errors, effectively avoids potential risks from unqualified partners, and improves the overall efficiency of cooperation and coordination.

[0105] Example 2:

[0106] This embodiment details the construction and execution process of the classification standard for processed parts, specifying the operational details of size classification, difficulty level determination, and classification execution.

[0107] I. Implementation of size classification standards;

[0108] Machine vision recognition equipment is used to inspect the dimensions of finished workpieces. The equipment has a camera resolution of 1920×1080 and the shooting angle is perpendicular to the surface of the workpiece. The Canny edge detection algorithm is used to extract the contour and then calculate the dimensions. The detection accuracy is ±0.01mm, which meets the accuracy requirements of the classification standard.

[0109] For plate parts, the length data of the finished product is obtained by taking pictures with the equipment. If the length is greater than 50mm, it is judged as a large part, and if the length is less than or equal to 50mm, it is a small or medium part. For rod parts, their diameter is measured. If the diameter is greater than 45mm, it is a large part, and if the diameter is less than or equal to 45mm, it is a small or medium part. For pipe parts, their outer diameter is measured. If the outer diameter is greater than 50mm, it is a large part, and if the outer diameter is less than or equal to 50mm, it is a small or medium part.

[0110] After the inspection is completed, the equipment automatically uploads the size data to the classification model, and the model outputs the size classification results in real time.

[0111] II. Determination of Difficulty Level Standards;

[0112] First, the hardness of the material is tested using a Rockwell hardness tester. The test points are selected on the key working surfaces of the machined parts. The average value of the three test points is taken as the final hardness value. The hardness is between HRC0 and HRC20, which is the easy level; between HRC20 and HRC40, which is the easy level; and greater than or equal to HRC40, which is the difficult level.

[0113] For stiffness determination, the length and outer diameter of the machined part are measured. The length is less than 500mm and the outer diameter is greater than 150mm, which is classified as simple. The length is less than 1000mm and 50mm less than the outer diameter and less than 100mm, which is classified as easy. The length is greater than 1000mm and 10mm less than the outer diameter and less than 50mm, which is classified as difficult.

[0114] The accuracy is determined based on the test results of the roughness tester. Ra100, Ra50, and Ra25 are classified as simple grades, Ra12.5, Ra6.3, and Ra3.2 as easy grades, and Ra1.6, Ra0.8, and Ra0.4 as difficult grades.

[0115] In tolerance assessment, free tolerance is classified as simple grade, ±0.05 as easy grade, and ±0.01 as difficult grade; parts with a semi-enclosed cavity or box-like shape are directly classified as difficult grade.

[0116] When a workpiece simultaneously meets multiple difficulty level characteristics, the most difficult level is used to determine the difficulty level. If a part has a hardness of HRC30 and a tolerance of ±0.01, it will be ultimately determined as a difficult level.

[0117] III. The construction process of the classification model is as follows:

[0118] We collected 500 sets of data on the dimensions, hardness, precision, tolerance, and shape of different types of machined parts, and divided them into training and testing sets in a 7:3 ratio. We used the support vector machine algorithm to build the model, setting the kernel function to the radial basis function, the penalty coefficient to 1.0, and the gamma value to 0.1.

[0119] The model parameters are optimized using the training set, and iterative training is stopped when the accuracy on the validation set reaches 99%. After the model is deployed, it can interact with machine vision recognition equipment, hardness testers, and roughness detectors in real time, automatically receiving detection data and outputting classification results.

[0120] IV. The calculation rules for the gross weight of materials are as follows:

[0121] For sheet metal, the gross weight of the material is calculated based on the maximum size of the processing drawing; for pipe, the gross weight of the material is calculated based on the maximum size of the processing drawing; and for bar, the gross weight of the material is calculated based on the next higher specification of the maximum size of the processing drawing.

[0122] In a certain processing drawing, the maximum size of the bar stock is φ40mm. The gross weight is calculated using a bar stock with a specification of φ45mm. This calculation rule is based on the industry standard for bar stock processing waste loss, ensuring the accuracy of material cost accounting.

[0123] This embodiment utilizes quantitative detection tools and clear classification rules, combined with a trained classification model, to achieve a precise and unified classification of workpiece dimensions and difficulty levels. This avoids the subjectivity of manual classification, provides a reliable basis for the scientific calculation of subsequent processing costs, and reduces disputes in the cost calculation process.

[0124] Example 3:

[0125] This embodiment details the specific operation process and parameter values ​​for determining the material-to-labor ratio coefficient, running the pricing model, and calculating costs.

[0126] The process for determining the material-to-labor ratio coefficient k is as follows:

[0127] We collected historical data on 100 different types and difficulty levels of machined parts, including information such as gross material weight, unit material price, processing difficulty characteristics, and actual processing costs. After standardizing the data, we used a linear regression algorithm to fit the initial value of the difficulty ratio k. The linear regression model formula is as follows: ;

[0128] in This is the ratio of actual processing costs to material costs. The comprehensive difficulty feature value, For regression coefficients, The intercept;

[0129] Five industry experts, each with over 10 years of experience in outsourced processing cost calculation, were invited to verify and adjust the initial values, ultimately determining the coefficient values: for large items with low difficulty... Easy Difficult Small and medium-sized items with low difficulty Easy Difficult ;

[0130] Collect 35 sets of new processing data each quarter, and use the gradient descent method to update the k value. Set the learning rate to 0.01 and the number of iterations to 100 to ensure that the coefficient can adapt to changes in industry processing technology and costs.

[0131] II. Profit Coefficient and tax coefficient The rules for determining the value are as follows:

[0132] Profit coefficient The initial value was set at 1.1, and adjusted to 1.08 based on the industry average profit margin changes over the past three years; tax coefficient. The initial value was 1.17, which was adjusted to 1.13 in accordance with national tax policies. The adjusted coefficient was verified by both the company's finance department and the industry association to ensure compliance and rationality.

[0133] III. The calculation of processing fees is divided into two scenarios: processing with materials included and processing with materials attached. The formula for processing fees with materials included is as follows: ;

[0134] The formula for processing costs with supplied materials is: ;

[0135] The meanings of each symbol are as follows: Total processing cost including materials (unit: yuan). Total cost of processing with materials (unit: yuan) Gross weight of materials (unit: kg). The price is the unit price of the material (unit: yuan / kg). This is the difficulty ratio. For profit coefficient, This is the tax coefficient. For shipping costs (unit: yuan).

[0136] IV. When the processed parts meet the requirements That is, when the length is ≥800mm or the diameter is ≥800mm, the material cost increases by 3% for the scrap ratio. This ratio is determined based on the measured data of the scrap loss rate in large part processing. The average scrap loss rate of 10 groups of large part processing was 2.8%, and 3% was taken as the calculation standard; the freight cost per batch The fixed price is 50 yuan, which is determined based on the average transportation cost of processed parts within the region. The transportation costs of 20 partners were statistically analyzed, and the average value was 48.5 yuan. 50 yuan was taken as the unified standard.

[0137] V. The pricing model construction and deployment process is as follows: The pricing model is built based on the Python language, integrating the above-mentioned cost calculation formula and coefficient update mechanism. It is encapsulated into an API interface through the Flask framework and deployed to a cloud server, supporting both batch data import and single-item calculation modes. When the model runs, it automatically reads the difficulty level output by the classification module to determine the k value. After receiving the gross weight and unit price data of materials, it calculates the cost in real time. The calculation results can be exported to Excel and PDF formats and are simultaneously synchronized to the contract module for contract generation.

[0138] Example calculation: A large part undergoing machining is made of 45 steel. What is the gross weight of the material? material unit price The difficulty level is Easy. Increased waste ratio and lower freight costs If the processing cost is 10 yuan, then the processing cost C = [(10×6)+(10×6×1.3)]×1.08×1.13+50 = [60+78]×1.2204+50≈168.42+50=218.42 yuan.

[0139] This embodiment determines the material-to-labor ratio coefficient through data fitting and expert verification. Combined with a dynamic adjustment mechanism and unified cost calculation rules, it constructs a scientific and standardized pricing model, ensuring the matching degree between processing costs and processing difficulty, improving cost transparency, and avoiding unreasonable pricing.

[0140] Example 4:

[0141] This embodiment details the specific implementation process of standardized contract signing, quality evaluation models, and full-cycle management mechanisms.

[0142] I. Standardized contracts will be signed using a one-year standardized contract template. The contract content will clearly define the rights and obligations of both parties, the price agreement, and the after-sales service method. The price agreement will clearly define the basis for calculating the processing fees, i.e., the relevant fee formulas and coefficient values. The after-sales service method will stipulate that the warranty period for the processed parts is one year, and the partner will replace the parts free of charge if any non-human-caused damage occurs during the warranty period. Before the contract is signed, it will be jointly reviewed by the company's legal and technical departments to ensure that the terms are free of legal risks and technical loopholes. The review approval rate must reach 100% before the contract can be signed. After the contract is signed, it will be completed online through the electronic signature system, and the contract data will be automatically synchronized to the system database for archiving.

[0143] II. The construction process of the quality assessment model is as follows:

[0144] Three indicators were selected: service life, machining accuracy, and failure rate. The weights were determined by the analytic hierarchy process, with service life having a weight of 0.4, machining accuracy having a weight of 0.3, and failure rate having a weight of 0.3.

[0145] The operational data of 200 processed parts were collected, including actual running time, accuracy deviation value, number of failures, etc. After the data was normalized, the fuzzy comprehensive evaluation method was used to train the model.

[0146] Setting factor set: ;

[0147] Collection of comments: ;

[0148] Construct a fuzzy evaluation matrix R, and obtain the evaluation result through the synthesis operation of the weight vector W and the evaluation matrix R. The synthesis operation formula is as follows:

[0149] Where B is the final evaluation vector. For fuzzy synthesis operators;

[0150] Each level corresponds to a specific threshold indicator. A five-star rating requires an operating life ≥ 120% of the agreed life, a machining accuracy deviation ≤ ±0.005mm, and a failure rate of 0. A product that is unusable requires an operating life < 50% of the agreed life, or a machining accuracy deviation > ±0.1mm, or a failure rate ≥ 3 times.

[0151] After the model is trained to a validation set with an accuracy of 97%, it is deployed to the full-cycle management module to receive the operation data of the processed parts in real time and output the evaluation results.

[0152] III. The implementation details of the quality responsibility mechanism are as follows:

[0153] When signing the contract, 10% of the total processing cost shall be reserved as a quality guarantee deposit, which shall be held in escrow by a third-party financial institution.

[0154] Assign dedicated personnel to handle daily communication with partners, track processing progress, and monitor quality control.

[0155] The technical department develops detailed technical standard documents that specify the material requirements, dimensional tolerances, surface treatment, and other details of the processed parts;

[0156] The commerce department oversees contract execution to ensure that partners deliver goods on time and to the agreed quality.

[0157] After the processed parts are installed, the operating data is collected in real time by sensors. The data transmission frequency is once per minute, and the evaluation results are output by the quality evaluation model.

[0158] If the equipment fails to reach the agreed operating period, the partner will replace it free of charge; if the number of failures exceeds three times, the entire quality deposit will be withheld and the cooperation will be terminated. At the same time, the partner will be blacklisted and no further cooperation will be carried out.

[0159] IV. The deployment of the full-cycle management and control process is as follows:

[0160] Establish a full-cycle management and control platform that integrates functional modules such as communication and tracking, technical standards management, contract supervision, and quality monitoring, with real-time data interaction between the modules;

[0161] The platform has an access control mechanism, with different departments having corresponding operating permissions to ensure data security; it generates a monthly quality control report to analyze the quality of processed parts and the performance of partners, providing data support for subsequent partner selection and technical standard optimization.

[0162] This embodiment clarifies the rights and obligations of both parties through standardized contracts, and relies on a multi-indicator quantitative evaluation model and quality responsibility mechanism to achieve effective full-cycle control of the quality of processed parts, standardize the performance of the partners, reduce contract disputes and quality problems, and improve the overall quality reliability of processed parts.

[0163] Example 5:

[0164] This embodiment focuses on system modules, detailing the specific implementation of each module and the data interaction process.

[0165] I. The partner management module uses an Oracle 19c database to store the partner directory and evaluation data. The database adopts a master-slave architecture to ensure high data availability and supports real-time data entry and query.

[0166] The module internally deploys a partner technical strength evaluation index system and risk assessment model. It receives real-time input partner data through a RESTful API interface, including the skill level of newly added operators and equipment update status. The model automatically calculates the technical strength score and risk level, and outputs a priority partner sequence. The data is updated once a day.

[0167] The module is deployed on a cloud server, using containerization technology to ensure deployment flexibility and supporting horizontal scaling to cope with data growth.

[0168] II. The machining parts classification module deploys machining parts classification standards and models. It receives parameters such as dimensions, material hardness, precision, tolerance, and shape of machining parts through a TCP / IP protocol interface. The dimensional data identified by machine vision and the hardness data detected by a hardness tester are automatically entered into the model. The model outputs the size classification and difficulty level classification results according to the classification rules. The classification results are output to the pricing module in real time through the interface. The module is equipped with a data verification mechanism to verify the format and numerical range of the input data. Data that does not meet the requirements returns an error message and requires re-uploading.

[0169] 3. The pricing module deploys the material cost ratio coefficient and pricing model, receives the classification results output by the processing parts classification module, and manually entered material gross weight m and material unit price p. The model automatically calculates the processing cost according to the cost formula. During the calculation process, it automatically determines whether the scrap ratio needs to be increased. If the condition is met, the material cost is automatically adjusted. The calculation results can be exported to Excel format. The module integrates a coefficient update interface, which automatically receives the updated material cost ratio coefficient, profit coefficient, and tax coefficient every quarter. The parameter update can be completed without manual intervention.

[0170] IV. The contract module stores standardized contract templates, supports contract generation and signing process management, and automatically calls the calculation results of the pricing module to fill in the price clauses when the contract is generated. After the contract is signed, the data is synchronized to the full-cycle management module. The module provides a contract template customization function, which can adjust the contract terms according to different processing types. It also supports a contract query function, which can search for contracts by keywords such as contract number and partner name.

[0171] V. The full-cycle management module deploys a processing part quality evaluation model, records the operation data of the processing parts, including service life, accuracy deviation, number of failures, manages quality assurance deposit information, updates the quality evaluation results of the processing parts in real time, and synchronizes the data to the query platform sub-module with a synchronization delay of no more than 5 minutes; the module is equipped with an early warning mechanism, which automatically sends an early warning notification to relevant management personnel when the processing parts have quality abnormalities or are close to the warranty period.

[0172] VI. The data interaction module connects to an external material price data platform, supporting automatic updates and imports of material unit prices, with an update frequency of once a week.

[0173] The contract information extraction unit constructs a regular expression matching library containing keywords related to manufacturing name, contractor, and signing date. This keyword library was determined through analysis of 100 outsourcing contract samples. After inputting an outsourcing contract, candidate information is first extracted using regular expression matching, and then a fuzzy inference model is used to calculate the truth value. The truth value calculation formula is as follows: ;

[0174] in For the first The truth value of each candidate piece of information. For the first The weights of the premise evidence range from [0.2, 0.5], set according to the importance of the evidence; the model outputs the key information with the highest truth value, and the key information extraction accuracy reaches over 95%; this unit is developed using Python and integrated into the data interaction module, supporting both batch processing and single contract processing modes.

[0175] VII. The query platform submodule is built using a B / S architecture. The front-end is developed using the Vue.js framework, and the back-end is developed using the Spring Boot framework. It displays the processing part classification results, partner unit rankings, processing cycles, and price prediction information. The data comes from the synchronized data of each functional module and is displayed after consistency verification. The verification rules are the correctness of the data format and the reasonableness of the numerical range. The data is updated once per hour to ensure the timeliness of the information. The platform supports access from multiple terminals, including computers, tablets, and mobile phones, adapting to different screen sizes. Access control is also set so that different users can view different ranges of information.

[0176] The system adopts a microservice architecture, with each module deployed independently and communicating through a service registration and discovery mechanism. Message queues are used to handle asynchronous data interactions to ensure stable system operation. The system has a logging module to record all operation logs and data interaction logs, facilitating troubleshooting and auditing. The system is regularly maintained and upgraded to ensure system security and performance stability.

[0177] This embodiment adopts a modular and information-based design, realizing real-time data interaction and collaborative work among various functional modules. Through automated information extraction and query functions, it reduces the redundancy of manual operations, improves the management efficiency and data accuracy of the entire outsourced processing process, and meets the needs of efficient control.

[0178] In summary, this outsourcing processing cost calculation method and dynamic management system based on the material-labor ratio allows the partner management module to pre-establish a categorized and ranked directory and a quantitative evaluation model to identify preferred partners. The processing parts classification module quickly outputs classification results through automated detection and classification models. The contract module provides standardized templates for rapid signing. The collaboration of these modules eliminates the need for scattered drawing review and bidding processes, meets the requirements for rapid response, and shortens the processing cycle.

[0179] Furthermore, the outsourcing processing cost calculation method and dynamic management system based on the material-labor ratio have a processing parts classification module that classifies parts by size and difficulty level according to unified standards, and a pricing module that automatically calculates costs based on the material-labor ratio coefficient, scientific formulas, and dynamically updated parameters, forming a unified pricing basis. This avoids arbitrary pricing by partners, allows companies to clearly control costs, and reduces their operating burden.

[0180] Furthermore, the outsourcing processing cost calculation method and dynamic management system based on the material-labor ratio, the partner management module quantitatively screens partners through technical strength and risk assessment models, the full-cycle control module deploys a quality evaluation model to collect data in real time to evaluate quality, and with the addition of quality assurance deposits and responsibility mechanisms, the drawbacks of subjective screening are eliminated, ensuring the quality of processed parts and after-sales service, and reducing the risk of safety accidents.

[0181] The relevant modules involved in this system are all hardware system modules or functional modules that combine computer software programs or protocols with hardware in the prior art. The computer software programs or protocols involved in these functional modules are technologies known to those skilled in the art and are not improvements to this system. The improvement of this system lies in the interaction or connection between the modules, that is, in improving the overall structure of the system to solve the corresponding technical problems that this system aims to address.

[0182] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for calculating outsourced processing costs based on the material-to-labor ratio, characterized in that, Includes the following steps: Step 1: Establish a directory of partner organizations and a contact information booklet, categorized and ranked by profession; Step 2: Construct a technical strength evaluation index system and risk assessment model for partners, conduct regular surveys on partners' technical strength, financial strength, and social reputation, and determine the priority partners through model scoring; Step 3: Establish classification standards for processed parts, classifying them into large and medium-sized parts according to size, and into three levels: simple, easy, and difficult according to material hardness, rigidity, precision, tolerance, and shape; Step 4: Determine the material-labor ratio coefficient, and calculate the processing cost using the coefficient based on the classification results. The formula is: ； The formula for processing costs with supplied materials is: ； in For the gross weight of the materials, This is the unit price of the material. The difficulty ratio and , The profit coefficient and , Tax coefficient and , For shipping costs; Step 5: Sign a one-year standardized contract, clearly defining rights and obligations, price agreement, and after-sales service methods; Step Six: Construct a quality evaluation model for processed parts, establish a quality responsibility mechanism for processed parts, reserve a quality guarantee deposit, and implement standardized management of the entire process through the division of labor of dedicated personnel for communication, technical standard formulation, and business plan supervision and implementation.

2. The method for calculating outsourced processing costs based on the material-labor ratio according to claim 1, characterized in that, The process for constructing the technical strength evaluation index system for partners, as described in step two, is as follows: Step 1: Select three core indicators: operator skill level, number of processing equipment, and establishment time; Step 2: Use the analytic hierarchy process (AHP) to determine the weights of the indicators: operator skill level weight 0.4, number of processing equipment weight 0.3, and establishment time weight 0.

3. Step 3: Collect historical data from no fewer than 50 partners. This sample size meets the minimum sample requirement of the random forest model after statistical testing. Divide the data into an 8:2 training set and a test set. Train the model using the random forest algorithm and iterate 500 times until the loss value is below 0.

05. Step 4: Ensure the model accuracy is no less than 85% through 5-fold cross-validation; Step 5: Input partner data in real time, and the model outputs a technical strength score: 5 points for technician, 3 points for senior worker, 2 points for intermediate worker, 1 point for junior worker, 5 points for more than 5 pieces of equipment, 1 point for less than 3 pieces of equipment, no points for no equipment, 5 points for establishment time of more than 5 years, 2 points for establishment time of less than 3 years, and no points for newly established companies.

3. The method for calculating outsourced processing costs based on the material-labor ratio according to claim 1, characterized in that, The process for constructing and implementing the workpiece classification criteria described in step three is as follows: Step 1: The size classification standard is determined based on the finished size of the processed parts. Plates with a length > 50mm are large parts, and those ≤ 50mm are small to medium parts. Rods with a diameter > 45mm are large parts, and those ≤ 45mm are small to medium parts. Pipes with an outer diameter > 50mm are large parts, and those ≤ 50mm are small to medium parts. The basis for using the corresponding classification standard for different types of parts is the difference in processing technology and industry processing practices. Step 2: The difficulty level standards are as follows: material hardness HRC0-HRC20 is easy, HRC20-HRC40 is easy, and ≥HRC40 is difficult; length <500mm and outer diameter >150mm is easy, length <1000mm and 50mm < outer diameter <100mm is easy, and length >1000mm and 10mm < outer diameter <50mm is difficult; Ra100, Ra50, Ra25 are easy, Ra12.5, Ra6.3, Ra3.2 are easy, and Ra1.6, Ra0.8, Ra0.4 are difficult; free tolerance is easy, ±0.05 is easy, and ±0.01 is difficult; semi-enclosed cavity box-type parts are classified as difficult. Step 3: When a workpiece simultaneously meets multiple difficulty level characteristics, the most difficult level shall be used to determine the difficulty level. Step 4: Identify the dimensions of the finished workpiece using machine vision, test the material hardness using a hardness tester, and input the results into the classification model to automatically output the classification results; Step 5: Calculate the gross weight of the bar stock according to the specification of the largest dimension above the maximum dimension in the processing drawing. The difference between the classification standard and the material calculation rule is based on the industry standard for waste loss in bar stock processing.

4. The method for calculating outsourced processing costs based on the material-labor ratio according to claim 1, characterized in that, The process for determining the material-to-labor ratio coefficient and operating the pricing model in step four is as follows: Step 1: Collect 100 sets of historical data on machined parts, and fit the difficulty ratio using a linear regression algorithm. Based on the verification and determination of coefficient values ​​by 5 industry experts, the difficulty of large items is simplified. Easy Difficult Small and medium-sized items are simple in difficulty Easy Difficult ; Step 2: Collect no fewer than 30 sets of new processing data each quarter, and update the coefficients and profit coefficients using the gradient descent method. Initially 1.1, later adjusted to 1.08, tax coefficient. The initial value was 1.17, which was later adjusted to 1.

13. The adjustment was based on national tax policies and changes in the industry's average profit margin. Step 3: Enter the gross weight of materials Material unit price Based on the classification results, the model automatically calculates the processing costs. Step 4: "≥L800(φ800)" means that the length of the processed part is ≥800mm or the diameter is ≥800mm. At this time, the material cost will increase by 3% of the scrap ratio, based on the actual measured data of the scrap loss rate of large parts processing. Step 5: Shipping costs per batch The amount is determined based on the average transportation cost of processed parts within the region.

5. The method for calculating outsourced processing costs based on the material-labor ratio according to claim 1, characterized in that, The process for constructing and operating the workpiece quality evaluation model described in step six is ​​as follows: Step 1: Select three indicators: service life, machining accuracy, and failure rate, with weights of 0.4, 0.3, and 0.3 respectively, and determine them using the analytic hierarchy process (AHP). Step 2: Collect operational data from 200 processed parts, train the model using the fuzzy comprehensive evaluation method, and set five levels from "five-star rating" to "product unusable", with each level corresponding to a specific indicator threshold; Step 3: Collect real-time data on the operation of the processed parts after installation, and output evaluation results from the model. If the parts are damaged before the agreed operating cycle is reached, the partner will replace them free of charge. If the damage exceeds three times, the deposit will be withheld and the cooperation will be terminated.

6. The method for calculating outsourced processing costs based on the material-labor ratio according to claim 1, characterized in that, The process for constructing and applying the risk assessment model for partners described in step two is as follows: Step 1: Construct a four-level evaluation index system, including project management risk, contractor management risk, project performance risk, and project quality risk. Each primary index contains 3-4 secondary indicators. Step 2: Invite 10 industry experts to score, use the analytic hierarchy process to determine the weights of each level of indicators, combine with the fuzzy comprehensive evaluation model for training, and iteratively optimize until the consistency test CR < 0.1; Step 3: The model outputs risk levels, including very high risk, high risk, medium risk, and low risk. Cooperation is possible for medium risk and below, and process monitoring should be strengthened. Cooperation is excluded for very high risk. The risk level classification is based on the risk occurrence probability and impact matrix.

7. A dynamic management system for outsourced processing costs based on the material-to-labor ratio, characterized in that, include: Partner Management Module: Stores partner directory and evaluation data, and deploys partner technical strength evaluation index system and risk assessment model; Processed Parts Classification Module: Deploys the processing parts classification standards and models, receives processing parts parameters and outputs classification results, which are then output to the pricing module as input parameters. Pricing module: Deploys the material cost ratio and pricing model, inputs the gross weight of materials, unit price of materials and classification results, and automatically calculates processing costs; Contract module: Stores standardized contract templates, supports contract generation and signing process management, and synchronizes contract data to the full lifecycle management module; Full-cycle management module: Deploys a processing part quality evaluation model, records processing part operation data, manages quality assurance deposits, and synchronizes data to the query platform sub-module; Data interaction module: Connects to external material price data, supports various data input and export, and includes a contract information extraction unit.

8. The dynamic management system for outsourced processing costs based on the material-labor ratio according to claim 7, characterized in that, The contract information extraction unit process of the data interaction module is as follows: Step 1: Build a regular expression matching library containing keywords for manufacturing name, contractor, and signing date. The keyword library was determined by analyzing 100 samples of outsourcing contracts. Step 2: Train the fuzzy inference filtering model. The truth value calculation formula is: in The weight of the premise evidence is set, with a value range of [0.2, 0.5], based on the importance of the evidence. Step 3: Input the outsourcing contract, first extract candidate information through regular expression matching, then use fuzzy reasoning to filter the model to calculate the truth value, and output the key information with the highest truth value.

9. A dynamic management system for outsourced processing costs based on the material-labor ratio according to claim 7, characterized in that, The query platform sub-module displays information such as processing part classification, partner unit ranking, processing cycle, and price forecast. The data is updated hourly to ensure information timeliness. The data comes from synchronized data from various functional modules and is displayed after consistency verification.

Images