Automatic proportioning and mixing system and method for cement pole raw materials based on order requirements

Abstract

The present invention discloses an automatic proportioning and mixing system and method for cement pole raw materials based on order requirements, and relates to the technical field of cement pole production. The method comprises the following steps: receiving a cement pole order and extracting key order information, forming an order feature vector according to the key order information, and splitting the order into multiple sub-orders; collecting historical production data of cement poles, and constructing a cement pole proportioning optimization model; traversing the sub-orders, generating proportioning variables and proportioning scheme candidate sets according to the key information of the sub-orders, and using the cement pole proportioning optimization model to obtain the material cost, strength performance and production efficiency prediction values ​​of each proportioning scheme, and screen out the optimal proportioning scheme. The method effectively reduces the cost of manual experience proportioning, improves the utilization rate of raw materials and the stability of product quality, and improves the production efficiency and economic benefits of the enterprise.

Description

Technical Field

[0001] The present invention relates to the technical field of cement pole production, and more specifically, to an automatic proportioning and mixing system and method for raw materials of cement poles based on order requirements. Background Art

[0002] At present, the raw material ratio in cement pole production mainly relies on manual experience, which is difficult to meet the needs of mass customized production, and the raw material utilization rate is low and the product quality is unstable. Although there are some methods for order optimization processing and multi-skilled worker matching in the prior art, there is no technical solution specifically for the automatic ratio of raw materials for cement poles.

[0003] Patent WO2024011844A1 discloses a method for optimizing the processing of aggregate orders, which generates multiple candidate aggregate orders using all orders in the order pool in a preset cycle, and adjusts the candidate aggregate orders that do not meet the preset constraints. Although this method can optimize the generation of orders, it does not involve the order production link and cannot achieve automatic matching of raw materials.

[0004] Patent CN108564244A discloses an order processing method, which obtains historical order information and processes it to obtain multiple order types, and then generates a matching model based on the mapping relationship between multi-skilled workers and order types. This method uses historical data to establish a matching model, which can reduce the impact of human factors, but its focus is on the matching of multi-skilled workers and order types, and also does not involve the issue of raw material ratio.

[0005] Therefore, how to realize automatic proportioning and mixing of raw materials in cement pole production in combination with order requirements, thereby reducing costs and improving production efficiency and product quality, is a technical problem that needs to be urgently solved in this field. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides an automatic proportioning and mixing system and method for cement pole raw materials based on order requirements.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The automatic proportioning and mixing method of cement pole raw materials based on order requirements includes:

[0009] Receive cement pole orders, extract key order information, and form an order feature vector based on the key order information; split the cement pole order into multiple sub-orders based on the order feature vector and a preset order splitting rule library to generate a sub-order set;

[0010] Collect historical production data of cement poles and build a cement pole ratio optimization model based on the historical production data;

[0011] Traverse the sub-order set, extract the key information of each sub-order, generate the ratio variable of the current sub-order according to the key information of the current sub-order; generate a candidate set of ratio schemes according to the ratio variables; obtain the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the candidate set of ratio schemes according to the key information of the current sub-order, the candidate set of ratio schemes and the cement pole ratio optimization model; screen out the optimal ratio scheme in the candidate set of ratio schemes according to the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the candidate set of ratio schemes.

[0012] Furthermore, the method for automatically proportioning and mixing raw materials of cement poles based on order requirements includes:

[0013] The order splitting rule base is expressed in the form of IF-THEN production. The IF part is the order feature vector, and the THEN part is the splitting rule corresponding to the order feature vector.

[0014] Automatically receive cement pole orders through the API interface of the order management system, use natural language processing technology to perform semantic analysis on the cement pole order text, and extract key order information;

[0015] Match the order feature vector with the IF part in the order splitting rule library to find all splitting rules applicable to the current order;

[0016] If multiple order splitting rules are matched, sort and filter them according to their priority and confidence, and select the optimal order splitting rule;

[0017] Call the optimal order splitting rule to split the cement pole order into multiple sub-orders;

[0018] Perform a legality check on the sub-orders to determine whether the sub-orders meet the requirements of the cement pole order. If they do not meet the requirements, adjust the order splitting rules and split the cement pole order again according to the adjusted order splitting rules. If they meet the requirements, continue to determine whether all sub-orders have undergone legality checks. If so, summarize all sub-orders to form a sub-order set; deduplicate and merge the sub-order set, and group and sort them according to the key order information; if not, perform a legality check on the next sub-order.

[0019] Furthermore, the method for automatically proportioning and mixing raw materials of cement poles based on order requirements includes:

[0020] Collect historical production data of cement poles, wherein the historical production data includes historical order demand data, historical production formula data, historical strength performance test data, historical material cost data and historical production efficiency data;

[0021] Clean, filter and extract features of historical production data to obtain order demand features, formula parameter features, strength performance features, material cost features and production efficiency features, and construct a sample data set based on order demand features, formula parameter features, strength performance features, material cost features and production efficiency features;

[0022] The cement pole ratio optimization model is a multi-objective optimization neural network model, which includes a backbone network and sub-networks. The sub-networks include a material cost sub-network, a strength performance sub-network and a production efficiency sub-network, which are used to predict material cost, strength performance and production efficiency respectively; the backbone network receives order demand features and formula parameter features, extracts high-dimensional features of order demand features and formula parameter features through a fully connected layer, and generates a high-dimensional feature vector; each sub-network receives the high-dimensional feature vector output by the backbone network, and maps it to the corresponding material cost prediction value, strength performance prediction value and production efficiency prediction value through a fully connected layer;

[0023] The loss functions of the material cost subnetwork, strength performance subnetwork and production efficiency subnetwork are designed respectively, and the loss function of the entire multi-objective optimization neural network model is designed according to the loss functions of the material cost subnetwork, strength performance subnetwork and production efficiency subnetwork.

[0024] Furthermore, the method for automatically proportioning and mixing raw materials of cement poles based on order requirements includes:

[0025] According to the key information of the current sub-order, the material selection and process matching of cement poles are carried out to obtain the raw material types and production processes;

[0026] Based on the key information, raw material types and production process of the current sub-order, determine the usage percentage range of each raw material for cement poles and generate the ratio variables of the current sub-order;

[0027] According to the material cost prediction value, strength performance prediction value and production efficiency prediction value of the mix ratio scheme, an evaluation point in the three-dimensional space of cost-strength-efficiency is constructed, and the comprehensive evaluation score of each mix ratio scheme is obtained according to the evaluation point;

[0028] According to the comprehensive evaluation scores of the matching schemes, the matching schemes in the matching scheme candidate set are sorted to obtain a matching scheme sequence from best to worst, which constitutes the optimal matching scheme set for the current sub-order;

[0029] Analyze the production requirements of cement pole products corresponding to the current sub-order, and adaptively select the optimal ratio scheme from the set of optimal ratio schemes;

[0030] Generate the cement pole production formula for the current sub-order based on the optimal ratio plan, add it to the production formula set of the cement pole order, and update the order progress information; generate cement pole production formulas for all sub-orders, and output the complete production formula set of this cement pole order.

[0031] Furthermore, the determination of the usage percentage range of each raw material of cement poles and the generation of ratio variables include:

[0032] Estimate the total volume based on the number, length and cross-sectional dimensions of the poles in the current sub-order, and convert the total amount of concrete; estimate the cement usage range per cubic meter of concrete based on the strength grade and production process; estimate the usage range of other materials by referring to the concrete mix design regulations; divide the usage range of each material by the total amount of concrete to obtain the usage percentage range of each raw material for cement poles, and generate the mix ratio variables of the current sub-order based on the usage percentage range of each raw material for cement poles;

[0033] An evaluation point in the three-dimensional space of construction cost-intensity-efficiency, and a comprehensive evaluation score of each ratio scheme obtained according to the evaluation point includes:

[0034] Divide the material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix ratio scheme in the mix ratio scheme candidate set by the maximum value of the corresponding prediction value to perform dimensionless processing on the indicators;

[0035] The dimensionless material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix ratio scheme are used as coordinates to mark the corresponding evaluation points in the three-dimensional space, and each evaluation point corresponds to a mix ratio scheme;

[0036] In the three-dimensional space composed of the material cost prediction value, the strength performance prediction value, and the production efficiency prediction value, the coordinates of the ideal optimal value point and the worst value point are calculated, where the optimal value of the material cost prediction value takes the minimum value and the worst value takes the maximum value; the optimal values ​​of the strength performance prediction value and the production efficiency prediction value take the maximum value and the worst value takes the minimum value; the coordinates of the optimal value point and the coordinates of the worst value point constitute the ideal point and the negative ideal point respectively;

[0037] Calculate the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point of each matching scheme; calculate the comprehensive evaluation score of the matching scheme according to the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point;

[0038] Methods for calculating the comprehensive evaluation score of the matching scheme based on the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point include:

[0039]

[0040] Among them, Si is the comprehensive evaluation score of the i-th matching scheme, is the Euclidean distance from the evaluation point to the ideal point of the i-th matching scheme, is the Euclidean distance from the evaluation point to the negative ideal point of the i-th matching scheme, and γ is an adjustment parameter used to control the sensitivity of the score.

[0041] Furthermore, the adaptively selecting the optimal matching scheme from the optimal matching scheme set includes:

[0042] Analyze the production requirements of cement pole products corresponding to the current sub-order, extract performance indicators, and establish a performance indicator set, including: performance indicator set KPI = {Q, R, S, E}, where Q is the raw material quality requirement, R is the ratio range, S is the strength grade requirement, and E is the production efficiency requirement;

[0043] For each performance indicator in the performance indicator set, the performance indicator threshold is set by comprehensively considering industry standards, customer requirements and enterprise quotas;

[0044] For the raw material quality requirement Q, set the raw material quality threshold Q0;

[0045] For the ratio range R, set the lower limit R of the amount of each raw material min And the upper limit of the dosage ratio R max ;

[0046] For strength level S, set the lower limit of strength index S min ;

[0047] For production efficiency E, set the upper limit of production time T max and energy consumption cap J max ;

[0048] Obtain the urgency and expected sales profit of the current sub-order, and dynamically adjust the performance indicator threshold according to the urgency and expected sales profit of the order;

[0049] Traverse the optimal ratio scheme set, determine whether each ratio scheme meets the dynamically adjusted performance indicator threshold conditions, and the schemes that meet the performance indicator threshold conditions enter the alternative scheme pool. When the ratio scheme does not meet the performance indicator threshold conditions, determine whether the alternative scheme pool is empty; when the alternative scheme pool is empty, adjust the performance indicator threshold, and re-determine whether the ratio scheme meets the adjusted performance indicator threshold conditions; when the alternative scheme pool is not empty, extract the material cost forecast value for the ratio scheme in the alternative scheme pool, and select the ratio scheme with the lowest cost as the optimal ratio scheme;

[0050] Further, the dynamically adjusting the performance indicator threshold according to the urgency of the order and the expected sales profit includes:

[0051] Let the urgency of the current sub-order be U, U∈[0,1]; let the expected sales profit be P, P∈[0,1];

[0052] When U>U0, the performance indicator threshold is adjusted as follows:

[0053] Q0=Q0+ΔQ;R min =R min -ΔR, R max =R max +ΔR;

[0054] S min =S min +ΔS;T max =T max -ΔT,J max =J max -ΔJ;

[0055] Among them, U0 is the preset urgency threshold; ΔQ is the preset adjustment step of the raw material quality threshold when U>U0; ΔR is the preset adjustment step of the lower limit and upper limit of the usage ratio when U>U0; ΔS is the preset adjustment step of the lower limit of the strength index when U>U0; ΔT is the preset adjustment step of the upper limit of the production time when U>U0; ΔJ is the preset adjustment step of the upper limit of energy consumption when U>U0;

[0056] When P>P0, the performance indicator threshold is adjusted as follows:

[0057] Q0=Q0-ΔQ′;R min =R min +ΔR′,R max =R max -ΔR′;

[0058] S min =S min -ΔS′;T max =T max +ΔT′,J max =J max +ΔJ′;

[0059] Among them, P0 is the preset sales profit threshold; ΔQ' is the preset adjustment step of the raw material quality threshold when P>P0; ΔR' is the preset adjustment step of the lower limit and upper limit of the usage ratio when P>P0; ΔS' is the preset adjustment step of the lower limit of the strength index when P>P0; ΔT' is the preset adjustment step of the upper limit of the production time when P>P0; ΔJ' is the preset adjustment step of the upper limit of energy consumption when P>P0.

[0060] The automatic proportioning and mixing system of cement pole raw materials based on order requirements is used to realize the automatic proportioning and mixing method of cement pole raw materials based on order requirements, including:

[0061] Order management module: used to receive cement pole orders, extract key order information and form order feature vectors, and verify and complete the order feature vectors;

[0062] Order splitting module: used to split the cement pole order into multiple sub-orders according to the order feature vector and the preset order splitting rule library, and generate a sub-order set;

[0063] Historical data collection module: used to collect historical order demand data, historical production formula data, historical strength performance test data, historical material cost data and historical production efficiency data of cement poles, and clean, filter and extract features from these data to build a sample data set;

[0064] Ratio design module: Design a cement pole ratio optimization model based on historical production data. The cement pole ratio optimization model is a multi-objective optimization neural network model, which includes a shared backbone network and three independent sub-networks. The backbone network extracts high-dimensional features of order demand characteristics and formula parameter characteristics, and outputs them to three sub-networks. Each sub-network predicts material cost, strength performance and production efficiency respectively. The loss function of each sub-network together constitutes the loss function of the entire multi-objective optimization neural network model.

[0065] Scheme screening module: used to traverse the sub-order set, extract the key information of each sub-order, generate the ratio variables of the current sub-order according to the key information of the current sub-order; generate a candidate set of ratio schemes according to the ratio variables; obtain the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the candidate set of ratio schemes according to the key information of the current sub-order, the candidate set of ratio schemes and the cement pole ratio optimization model; according to the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the candidate set of ratio schemes, screen out the optimal ratio scheme in the candidate set of ratio schemes.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] The present invention realizes automatic order reception and key information extraction by introducing automated order management and processing procedures, avoids the inefficiency and error-proneness of manual processing, and greatly improves the speed and accuracy of order processing. The present invention uses historical production data to establish a cement pole ratio optimization model. By traversing sub-orders and generating a candidate set of ratio schemes, the material cost, strength performance and production efficiency of each ratio scheme can be accurately predicted, thereby screening out the optimal ratio scheme. This not only reduces the waste of raw materials and improves material utilization, but also significantly improves the quality stability and consistency of the product. In addition, the application of the multi-objective optimization neural network model enables the system to optimize in three dimensions of material cost, strength performance and production efficiency at the same time, ensuring that the best production performance and economic benefits are achieved while meeting order requirements. By matching the order feature vector and the split rule library, the system can flexibly respond to orders of different types and sizes, and improve the flexibility and response speed of production scheduling. Overall, the present invention greatly improves the intelligence and automation level of cement pole production, reduces the dependence on manual experience, promotes the improvement of enterprise production efficiency and economic benefits, and enables enterprises to have stronger competitiveness and adaptability in the fierce market competition. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0069] Figure 1 It is a flow chart of the method for automatically proportioning and mixing raw materials of cement poles based on order requirements in the present invention;

[0070] Figure 2 It is a flow chart of a method for obtaining a comprehensive evaluation score of a proportioning scheme in a method for automatically proportioning and mixing raw materials for cement poles based on order requirements of the present invention;

[0071] Figure 3 It is a functional module diagram of the automatic proportioning and mixing system of cement pole raw materials based on order requirements in the present invention. DETAILED DESCRIPTION

[0072] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0073] Example 1

[0074] See also Figure 1 As shown, this embodiment provides an automatic proportioning and mixing method for cement pole raw materials based on order requirements, including:

[0075] Step S1000, receiving a cement pole order, extracting key order information, and forming an order feature vector according to the key order information; splitting the cement pole order into multiple sub-orders according to the order feature vector and a preset order splitting rule library, and generating a sub-order set;

[0076] Furthermore, step S1000 includes:

[0077] Step S1100, automatically receiving cement pole orders through the API interface of the order management system, extracting key order information, and forming an order feature vector based on the key order information;

[0078] Further, step S1100 includes the steps of:

[0079] Step S1110, automatically receiving cement pole orders through the API interface of the order management system, using natural language processing technology to perform semantic analysis on the cement pole order text, and extracting key order information; the key order information includes pole type, pole specification, pole quantity, strength grade and delivery time;

[0080] Step S1120, structure the key order information to form a standardized order feature vector;

[0081] Step S1130, verifying and completing the order feature vector;

[0082] Specifically, cement pole orders are automatically obtained through the API interface of the order management system. These orders include pole demand information submitted by customers. The system pulls and receives this data through a scheduled task or event trigger mechanism to ensure the real-time and completeness of the order information. The received order text is semantically parsed using natural language processing technology. The system uses a well-trained natural language processing model to identify key information in the order, such as pole type, pole specification, pole quantity, strength level, delivery time, etc. The semantic parsing process includes steps such as word segmentation, part-of-speech tagging, and named entity recognition to ensure the accuracy of the extracted information. The extracted key information is structured according to the predefined data model to form a standardized order feature vector. The data model defines the format and type of each order attribute to ensure that all order information is represented in a consistent manner. The structured representation enables the order data to be efficiently processed and used by other modules in the system. The order feature vector is verified and completed to ensure the integrity and accuracy of the data. The verification process includes checking the data format, range, and logical consistency. For missing or incomplete information, the system completes it with historical data or default values ​​to ensure the integrity of the order data.

[0083] Step S1100 automatically obtains cement pole orders and extracts key information to form a standardized order feature vector, thereby realizing automation and intelligence of order processing, reducing manual workload and improving efficiency and accuracy of order processing.

[0084] Step S1200, loading a preset order splitting rule library, and reading predefined order splitting rules from the order splitting rule library; the order splitting rule library is expressed in the form of IF-THEN production, where the IF part is the order feature vector, and the THEN part is the splitting rule corresponding to the order feature vector;

[0085] Specifically, the order splitting rules are summarized and refined by industry experts based on years of experience, and include the optimal splitting rules for various typical order scenarios; the order splitting rule library uses the IF-THEN form of production expression. Among them, the IF part is the order feature vector, which contains the specific attributes and conditions of the order; the THEN part is the corresponding splitting rule, which defines how to split the order when specific conditions are met. This representation method is concise and clear, and is easy for system analysis and execution.

[0086] Step S1300, based on the order feature vector, the optimal order splitting rule for the cement pole order is screened out from the order splitting rule library; the optimal order splitting rule is called to split the cement pole order into multiple sub-orders to generate a sub-order set;

[0087] Furthermore, step S1300 includes:

[0088] Step S1310, matching the order feature vector with the IF part in the order splitting rule library to find out all splitting rules applicable to the current order;

[0089] Step S1320: if multiple order splitting rules are matched, sort and filter the order splitting rules according to their priorities and confidences, and select the optimal order splitting rule;

[0090] Step S1330, calling the optimal order splitting rule to split the cement pole order into multiple sub-orders;

[0091] Step S1340, checking the legitimacy of the sub-order to determine whether the sub-order meets the requirements of the cement pole order. If so, proceed to step S1350. If not, adjust the order splitting rules and re-split the cement pole order according to the adjusted order splitting rules.

[0092] Step S1350, determining whether all sub-orders have been subjected to a legality check, if so, executing step S1360, if not, repeating steps S1340 to S1350 to perform a legality check on the next sub-order;

[0093] Step S1360: Aggregate all sub-orders to form a sub-order set; remove duplicates and merge the sub-order set, and group and sort them according to key order information;

[0094] Specifically, the pattern matching algorithm is used to match the order feature vector with the IF part in the order splitting rule library. The system finds all splitting rules applicable to the current order based on the key order information in the order feature vector. The most relevant rules are screened out by calculating similarity or other matching indicators. In the case of matching multiple rules, each rule will have a preset priority. The priority is set by industry experts based on the importance and application scenarios of the rules. The system will sort the rules according to the priority, and the rules with higher priorities will be considered first. In addition to the priority, each rule will also have a confidence. This is an indicator evaluated based on historical data and the effect of rule application, reflecting the reliability and applicability of the rule. The higher the confidence, the more reliable the rule. The system sorts and screens according to the priority and confidence of the rules. Rules with higher priority and greater confidence are selected first. The system selects the optimal splitting rule by comprehensively considering multiple factors to ensure the rationality and effectiveness of the splitting rule. The optimal order splitting rule is called to split the cement pole order into multiple sub-orders; this order splitting rule ensures that the pole specifications and strength grades within each sub-order are consistent to the greatest extent, thereby simplifying the subsequent production and ratio optimization process. The split results are checked for legality to ensure that the split sub-orders meet the requirements of the cement pole order in terms of quantity, specifications, etc. The system verifies the split sub-orders one by one through preset verification rules and logic to ensure that the attributes and quantity of each sub-order are consistent with the requirements of the original order. If problems are found, the system will automatically adjust the splitting strategy and re-split until all sub-orders are legal and valid. Deduplication and merging of sub-order sets are to ensure that each sub-order is an independent, minimum-granularity unit to be optimized; the sub-order sets are grouped and sorted according to the key information of the order to facilitate subsequent batch optimization processing.

[0095] For example, suppose an order for cement poles is received, requiring the production of 100 8-meter high-strength poles and 200 10-meter ordinary poles, with delivery dates of 1 month and 2 months respectively. The system first performs semantic analysis on the order, extracts key attributes such as pole specifications (8 meters / 10 meters), quantity (100 poles / 200 poles), strength level (high strength / ordinary), and delivery date (1 month / 2 months), and forms a structured order feature vector. Then, the preset order splitting rule library is loaded, and it is found that one of the rules applies to the order: IF (the order contains poles of different specifications) AND (the corresponding delivery dates are inconsistent) THEN (split the order by specification and delivery date). According to this rule, the system splits the original order into two sub-orders: Sub-order 1 contains 100 8-meter high-strength poles with a delivery date of 1 month; Sub-order 2 contains 200 10-meter ordinary poles with a delivery date of 2 months. The sub-orders after splitting have consistent specifications and unified delivery dates, and can enter the subsequent ratio optimization link separately, which not only reduces the constraints during optimization, but also makes it easier to carry out batch production, thereby improving production efficiency and timeliness of delivery.

[0096] Step S1300 realizes efficient processing of cement pole orders through a systematic and automated order splitting method, significantly improving the accuracy and rationality of order splitting. Natural language processing technology is used to accurately extract key order information, and intelligent splitting is performed based on the preset splitting rule library. Through comprehensive sorting of priority and confidence, the selected splitting rule is ensured to be optimal, so that the generated sub-orders not only meet the needs of the original order, but also simplify the subsequent production and ratio optimization process to the greatest extent. The legality check and automatic adjustment mechanism further ensure the effectiveness and rationality of the sub-orders, reduce manual intervention, and improve production efficiency and the timeliness of order delivery. Overall, this method makes up for the inefficiency and high error rate of traditional manual splitting, and is particularly suitable for the processing of large-volume, multi-variety complex orders. It has significant advantages in improving the production management level of enterprises and optimizing resource allocation.

[0097] Step S1000 receives cement pole orders, extracts key order information and forms order feature vectors, and splits the orders into multiple sub-orders according to the preset order splitting rule base, which significantly improves the efficiency and accuracy of order processing. The use of automation and intelligent technology reduces the burden of manual processing and the probability of errors, ensuring the integrity and consistency of order data. Through natural language processing technology and structured representation, the system can accurately extract order information, form standardized order feature vectors, and perform intelligent splitting according to the optimal splitting rules. The generated sub-order set not only meets the original order requirements, but also simplifies the subsequent production and ratio optimization process, improving the company's production management efficiency and the timeliness of order delivery.

[0098] Step S2000, collecting historical production data of cement poles, and constructing a cement pole ratio optimization model based on the historical production data;

[0099] Furthermore, step S2000 includes:

[0100] Step S2100, collecting historical production data of cement poles, cleaning, screening and feature extraction of the historical production data, and constructing a sample data set;

[0101] Furthermore, step S2100 includes:

[0102] Step S2110, collecting historical production data of cement poles, wherein the historical production data includes historical order demand data, historical production formula data, historical strength performance test data, historical material cost data and historical production efficiency data;

[0103] Specifically, the historical order demand data reflects the various types of cement pole order information received by the company in the past, including key demand information such as pole specifications, quantity, and strength grade. The historical production formula data comes from the company's actual production records over the past period of time, including the types of raw materials used in each formula, the proportion of the mix, and the production process parameters. The historical strength performance test data includes the test results of a series of mechanical performance indicators such as compressive strength, flexural strength, crack resistance, freeze-thaw cycles, and aging tests, which can fully reflect the long-term performance of pole products. The historical material cost data records in detail the purchase price and actual usage of various raw materials required for production, as well as energy consumption, labor expenses, transportation costs, and other cost information during the production process, providing data reference for cost optimization. In addition, the historical production efficiency data includes production cycle, equipment utilization, energy consumption, labor hours, and fault downtime records, reflecting the actual operating efficiency of the company's production line.

[0104] Step S2110 collects historical production data of cement poles and constructs a sample data set, which includes multi-dimensional data such as historical order requirements, production formulas, strength performance, material costs, and production efficiency, providing a solid data foundation for subsequent ratio optimization work. This step fully taps into the valuable production and operation data accumulated by the company over the years, avoids data waste and loss, and preserves past experience and knowledge in a structured form. Using these historical data, the ratio optimization model can more accurately learn the mapping relationship between ratio and performance, shorten the cycle of new formula development, improve optimization accuracy, and make the optimization results more reliable and closer to reality.

[0105] Step S2120, cleaning, screening and feature extraction of historical production data to obtain order demand features, formula parameter features, strength performance features, material cost features and production efficiency features, and constructing a sample data set based on the order demand features, formula parameter features, strength performance features, material cost features and production efficiency features;

[0106] Specifically, the data preprocessing operation in step S2120 can improve sample quality, reduce the interference of noise and abnormal data, and convert the original data into structured features that can be directly used by the machine learning algorithm to prepare for subsequent model training. Through feature selection and feature engineering, redundant and irrelevant information in the data can be removed, key factors that have a greater impact on the ratio optimization task can be highlighted, and the efficiency and performance of model training can be improved. Data preprocessing is an important part of the machine learning process, and good data quality is the basis of model optimization.

[0107] Step S2200, designing a cement pole ratio optimization model structure according to the sample data set;

[0108] Further, step S2200 includes:

[0109] Step S2210, the cement pole ratio optimization model is a multi-objective optimization neural network model, which includes a shared backbone network and three independent sub-networks, wherein the three independent sub-networks are a material cost sub-network, a strength performance sub-network and a production efficiency sub-network, which are used to predict material cost, strength performance and production efficiency respectively; the backbone network receives order demand features and formula parameter features, extracts high-dimensional features of order demand features and formula parameter features through several fully connected layers, generates high-dimensional feature vectors, and outputs the extracted high-dimensional feature vectors to the three sub-networks; each sub-network receives the high-dimensional feature vector output by the backbone network, and maps it to the corresponding material cost prediction value, strength performance prediction value and production efficiency prediction value through one or two fully connected layers;

[0110] Step S2220, respectively designing the loss functions of the material cost subnetwork, the strength performance subnetwork, and the production efficiency subnetwork, and then designing the loss function of the entire multi-objective optimization neural network model based on the loss functions of the material cost subnetwork, the strength performance subnetwork, and the production efficiency subnetwork;

[0111] Assume that the multi-objective optimization neural network model has m samples in a training batch, and the true material cost, true strength performance and true production efficiency of the i-th sample are The material cost prediction value, strength performance prediction value and production efficiency prediction value of the model are

[0112] The calculation method of the material cost sub-network loss function includes:

[0113]

[0114] in:

[0115] L cost : Error loss function of the material cost subnetwork;

[0116] w c : The weight coefficient of material cost loss is determined by a large number of experiments by technicians in this field;

[0117] Pc: L cost The power exponent of the error loss function. By adjusting Pc, the sensitivity of the loss function to outliers can be controlled.

[0118] The relative error compares the difference between the predicted value and the true value with the true value itself, rather than simply considering the absolute difference. This is more in line with the relative importance of costs in actual situations, because the same absolute difference may have different effects at different cost levels. c The sensitivity of the loss function to cost errors can be adjusted.

[0119] The calculation method of the strength performance sub-network loss function includes:

[0120]

[0121] in:

[0122] L strength : Error loss function of the strength performance sub-network;

[0123] w s : The weight coefficient of mechanical property loss is determined by a large number of experiments by technicians in this field;

[0124] α: Penalty factor, α>1, used to amplify the loss when the strength is underestimated.

[0125] The piecewise loss function uses a higher penalty when the predicted value is lower than the true value, which helps the model pay more attention to the accuracy of mechanical properties, especially to prevent the occurrence of underestimation errors. By increasing the penalty for underestimation errors, the piecewise loss function can improve the reliability of the model for strength properties and ensure that the model's predicted results are closer to the true value.

[0126] The calculation method of the production efficiency sub-network loss function includes:

[0127]

[0128]

[0129] in:

[0130] Lprod: loss function of the production efficiency sub-network;

[0131] w p : The weight coefficient of production efficiency loss is determined by a large number of experiments by those skilled in the art;

[0132] β: a hyperparameter that controls the smooth segment of the loss function;

[0133] Smooth L1 loss function;

[0134] Apply the smooth L1 loss function to the error between the predicted value and the true value;

[0135] The production efficiency sub-network loss function has good numerical stability when the error is small, avoiding the problem of gradient explosion or vanishing; it has a stronger gradient when the error is large, which helps the model learn faster and adjust production efficiency.

[0136] The method of designing the loss function of the entire multi-objective optimization neural network model according to the loss functions of the material cost sub-network, the strength performance sub-network and the production efficiency sub-network includes:

[0137] L total =r c log(1+L cost )+r s log(1+L strength )+r p log(1+L prod );

[0138] in:

[0139] r c : Weight coefficient of material cost loss;

[0140] r s : Weight coefficient of strength performance loss;

[0141] r p : Weight coefficient of production efficiency loss;

[0142] L total : Loss function of multi-objective optimization neural network model;

[0143] This formula can adaptively reduce the difference between losses of different dimensions through logarithmic transformation while maintaining their original relative proportional relationship. c 、r s and r p It is determined by a large number of experiments by those skilled in the art that by adjusting rc 、r s and r p , the importance of each sub-goal can be flexibly controlled. This is necessary for multi-objective optimization because there may be trade-offs between different objectives.

[0144] Specifically, step S2200 constructs a machine learning model for multi-objective optimization through the designed neural network structure and loss function. The shared backbone network can extract the general feature representation of order and ratio information to avoid repeated calculations; the independent sub-networks are responsible for completing the prediction tasks of their respective optimization objectives, and the structure can be flexibly customized according to the complexity of different objectives. The weighted summation of the loss functions of the three sub-objectives can weigh the contribution of different indicators to the comprehensive optimization and guide the model to learn and optimize in the direction of taking into account multiple objectives. Dimensionless processing can eliminate the influence of errors of different dimensions and make the multi-objective losses more balanced in numerical terms.

[0145] Step S2300, using the sample data set to train and optimize the cement pole ratio optimization model;

[0146] The method of training and optimizing the cement pole ratio optimization model using the sample data set includes:

[0147] The sample data set is divided into a training set and a test set, and various hyperparameters for model training are set. Using the training set data, the parameters of each layer of the model are continuously updated through the back propagation algorithm and the gradient descent method, so that it can gradually fit the training samples and learn the mapping law of the optimized ratio scheme. According to the current model parameters and the characteristics of a batch of training samples, the material cost prediction value, strength performance prediction value and production efficiency prediction value are calculated through forward propagation; the prediction error is calculated based on the material cost prediction value, strength performance prediction value and production efficiency prediction value predicted by the current batch and the actual material cost, actual strength performance and actual production efficiency; based on the prediction error, the parameters of each layer of the model are updated by gradient descent; and the training data are iterated for multiple rounds until the performance of the model on the test set meets expectations.

[0148] Step S2300 uses the sample data set to train and optimize the cement pole ratio optimization model, significantly improving the scientificity and rationality of the ratio scheme. The trained and optimized model can not only meet the specific needs of different orders, but also minimize material costs and improve production efficiency while ensuring product quality. This process has a high degree of automation, reduces human intervention and errors, and improves the level of intelligent production management. Overall, step S2300 significantly improves the resource utilization and economic benefits in the cement pole production process through data-driven scientific methods, providing strong technical support for improving the competitiveness of enterprises.

[0149] Step S3000, traverse the sub-order set, extract the key information of each sub-order, generate the ratio variable of the current sub-order according to the key information of the current sub-order; generate a ratio scheme candidate set according to the ratio variable; obtain the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the ratio scheme candidate set according to the key information of the current sub-order, the ratio scheme candidate set and the cement pole ratio optimization model; screen out the optimal ratio scheme in the ratio scheme candidate set according to the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the ratio scheme candidate set;

[0150] Furthermore, step S3000 includes:

[0151] Step S3100, traverse the sub-order set, extract the key information of the current sub-order, determine the usage percentage range of each raw material of the cement pole according to the key information of the current sub-order, and generate the ratio variable of the current sub-order; within the value range of each ratio variable, gradually sample with a preset step length, list all possible ratio schemes, and form a candidate set of ratio schemes;

[0152] Further, step S3100 includes:

[0153] Step S3110, traverse the sub-order set and extract the key information of the current sub-order; according to the key information of the current sub-order, perform material selection and process matching for cement poles to obtain the type of raw materials and production process; the key information of the current sub-order includes pole strength grade, pole quantity, length, cross-sectional size, delivery date and special performance requirements; the special performance requirements include frost resistance, durability and impermeability;

[0154] First, key information related to the proportion design is extracted from the current sub-order information, mainly including: 1) Pole strength grade: such as C65, C80, etc., which directly affects the proportion design strength target of the pole; 2) Pole length and cross-sectional size: determine the volume and material amount of the pole, and are closely related to the total material consumption; 3) Pole quantity and delivery date: affect the economy and production efficiency requirements of the proportion scheme; 4) Other special performance requirements: such as frost resistance, durability, impermeability, etc., have specific requirements for material selection and proportioning.

[0155] According to the specific requirements of the sub-order for the performance of the pole product, select the matching raw material type, production process and quality grade from the material library, such as: 1) Select the cement grade according to the strength grade requirement: ordinary Portland cement P.O42.5, P.O52.5, etc.; 2) Select admixtures according to the pole production process: fly ash is often added in the autoclave curing method, and mineral powder is often added in the natural curing method; 3) Select admixtures according to durability requirements: high-performance water reducer, air entraining agent, etc.; 4) Select sand and gravel according to aggregate quality requirements: washed sand, continuously graded crushed stone, etc. The material selection process needs to follow professional mix design specifications, such as the national standard "GB / T 31387-2015 Concrete Mix Design Code for Prestressed Concrete Poles", etc., and also needs to be combined with the company's production process flow chart and material inventory, and completed by experienced mix experts or knowledge engine systems.

[0156] Step S3120, based on the key information, raw material types and production process of the current sub-order, determine the usage percentage range of each raw material of the cement pole, and generate the ratio variable of the current sub-order;

[0157] The methods for determining the usage percentage range of each raw material for cement poles and generating ratio variables include:

[0158] Estimate the total volume based on the number, length and cross-sectional dimensions of the poles in the current sub-order, and convert the total amount of concrete; estimate the cement usage range per cubic meter of concrete based on the strength grade and production process; estimate the usage range of other materials by referring to the concrete mix design regulations; divide the usage range of each material by the total amount of concrete to obtain the usage percentage range of each raw material for cement poles, and generate the mix ratio variables of the current sub-order based on the usage percentage range of each raw material for cement poles;

[0159] For example, the cement dosage range per cubic meter of concrete can be estimated based on the strength grade and production process. For example, the general dosage of C80 grade autoclave-cured poles is 400-450kg / m 3 ; Refer to the concrete mix design regulations to estimate the usage range of other materials, such as the usage range of fly ash is generally 100-150kg / m 3 , sand rate is controlled at 35%-45%, etc.; the usage range of each material is divided by the total usage of concrete to obtain the usage percentage range of each raw material of cement poles, for example: cement variable: [18%, 22%]; sand and gravel variable: [52%, 60%]; fly ash variable: [8%, 12%]; mineral powder variable: [6%, 10%]; water reducer variable: [1.0%, 1.8%]; water variable: [8%, 12%].

[0160] Step S3130, within the value range of each matching variable, sampling step by step with a preset step length, listing all possible matching schemes to form a matching scheme candidate set;

[0161] For example, cement variables are sampled in the range of [18%, 22%] with a step size of 1%, sand and gravel variables are sampled in the range of [52%, 60%] with a step size of 2%, and water reducers are sampled in the range of [1.0%, 1.8%] with a step size of 0.1%. The resulting candidate set of mix ratios may contain tens of thousands of mix ratios. Exhaustive sampling to generate candidate sets of mix ratios does not miss any possible high-quality mix ratios, which is convenient for parallel computing.

[0162] Step S3100 ensures that the key information of each sub-order is accurately and comprehensively collected and processed through a systematic traversal and extraction process, thereby providing a reliable data basis for subsequent ratio optimization. By determining the percentage range of raw material usage and generating ratio variables, it can flexibly adapt to the specific needs of different orders, increasing the diversity and optionality of ratio schemes. Gradually sampling and listing all possible ratio schemes to form a candidate set of ratio schemes helps to fully cover all potential ratio combinations, thereby improving the reliability and effectiveness of the final optimization results. This process not only significantly reduces the need for manual intervention, but also improves the efficiency and accuracy of order processing through automation and intelligent means, ensuring the smooth progress of the production process and the timely delivery of orders.

[0163] Step S3200, extracting the order demand characteristics of the current sub-order and the formula parameter characteristics of each mix scheme in the mix scheme candidate set, inputting the order demand characteristics and the formula parameter characteristics into the cement pole mix optimization model, and obtaining the material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix scheme in the mix scheme candidate set;

[0164] Step S3300, constructing an evaluation point in the cost-strength-efficiency three-dimensional space according to the material cost prediction value, strength performance prediction value and production efficiency prediction value of the mix ratio scheme, and obtaining a comprehensive evaluation score for each mix ratio scheme according to the evaluation point;

[0165] For further information, see Figure 2 , step S3300 includes:

[0166] Step S3310, dividing the material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix ratio scheme in the mix ratio scheme candidate set by the maximum value of the corresponding prediction value, and performing dimensionless processing on the indicators;

[0167] The maximum values ​​of the corresponding predicted values ​​refer to the maximum value of the material cost predicted values, the maximum value of the strength performance predicted values, and the maximum value of the production efficiency predicted values ​​of all the mix ratio schemes in the candidate set of mix ratio schemes; through dimensionless processing, the predicted values ​​of different dimensions and numerical ranges are unified into the [0,1] interval, eliminating the dimensional influence between the predicted values, which is convenient for subsequent comprehensive evaluation.

[0168] Step S3320, using the dimensionless material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix ratio scheme as coordinates, marking corresponding evaluation points in the three-dimensional space, and each evaluation point uniquely corresponds to a mix ratio scheme;

[0169] Step S3330, in the three-dimensional space formed by the material cost prediction value, the strength performance prediction value, and the production efficiency prediction value, the coordinates of the ideal optimal value point and the worst value point are calculated, wherein the optimal value of the material cost prediction value takes the minimum value, and the worst value takes the maximum value; the optimal values ​​of the strength performance prediction value and the production efficiency prediction value take the maximum value, and the worst values ​​take the minimum value; the coordinates of the optimal value point and the coordinates of the worst value point constitute an ideal point and a negative ideal point respectively;

[0170] Step S3340, calculating the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point of each matching scheme; calculating the comprehensive evaluation score of the matching scheme according to the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point;

[0171] Methods for calculating the comprehensive evaluation score of the matching scheme based on the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point include:

[0172]

[0173] in:

[0174] S i : Comprehensive evaluation score of the i-th matching scheme;

[0175] The Euclidean distance from the evaluation point to the ideal point of the i-th matching scheme;

[0176] The Euclidean distance from the evaluation point of the i-th matching scheme to the negative ideal point;

[0177] γ: adjustment parameter used to control the sensitivity of the score, usually between 1 and 2;

[0178] When the evaluation point is close to the ideal point:

[0179] Smaller, Larger; Close to 0, resulting in Also close to 0, the overall evaluation score is S i Close to 1;

[0180] When the evaluation point is close to the negative ideal point:

[0181] Larger, Smaller; Approaching infinity, leading to It is also close to infinity, and the comprehensive evaluation score is S i Close to 0;

[0182] By calculating the distance from each ratio scheme to the ideal point and the negative ideal point, the quality of the scheme can be quantitatively characterized. The closer the scheme is to the ideal point and the farther it is from the negative ideal point, the better its overall performance. The higher the comprehensive evaluation score, the farther the ratio scheme is from the negative ideal point, and the closer it is to the ideal point, the better its overall performance. The comprehensive evaluation score quantifies the evaluation results of each ratio scheme into a score in the [0,1] interval. The higher the score, the higher the cost-effectiveness of the scheme. The optimal ratio scheme can be quickly screened out by sorting the scores.

[0183] Step S3400: sort the matching schemes in the matching scheme candidate set according to the comprehensive evaluation scores of the matching schemes to obtain a matching scheme sequence from best to worst, which constitutes the optimal matching scheme set for the current sub-order;

[0184] Through comprehensive evaluation and sorting, a group of optimal ratio schemes that take into account cost, strength and efficiency are selected from the candidate set of ratio schemes, providing a high-quality initial solution for subsequent production formula optimization.

[0185] Step S3200, step S3300 and step S3400 can achieve the best balance between cost, strength and efficiency and comprehensively improve the economic efficiency and product quality of production by optimizing in the three-dimensional space composed of the three dimensions of material cost prediction value, strength performance prediction value and production efficiency prediction value. The Euclidean distance calculation method of ideal point and negative ideal point is used to accurately and comprehensively evaluate each ratio scheme. This method quantifies the advantages and disadvantages of different ratio schemes, provides an objective evaluation standard, and avoids the deviation of subjective judgment. Taking the minimum value of the optimal value of the material cost prediction value ensures that the production cost is minimized while ensuring product quality and production efficiency, which helps enterprises maintain cost advantages in market competition. Taking the maximum value of the optimal value of the strength performance prediction value and the production efficiency prediction value ensures that the produced cement poles reach the best state in terms of strength and efficiency, and improves the market competitiveness and customer satisfaction of the product. Based on accurate prediction models and historical data, the evaluation and selection are carried out by calculating the Euclidean distance to ensure that the decision-making process is data-driven, reduce the interference of human factors, and improve the scientificity and reliability of decision-making. This method not only focuses on the optimization of a single indicator, but also evaluates from multiple dimensions through comprehensive consideration of three-dimensional space, which improves the comprehensiveness and scientificity of the selection of the ratio scheme. The functions of real-time calculation and dynamic adjustment enable it to flexibly select the optimal ratio scheme according to the specific needs and production conditions of different orders, enhancing the flexibility and adaptability of the production process. By quickly screening out the optimal ratio scheme, a large number of trial and error and verification processes in traditional methods are reduced, greatly improving production efficiency. This method ensures that enterprises can find the optimal ratio scheme in the shortest time, thereby improving production efficiency and product quality, reducing production costs, and enhancing the competitiveness of enterprises in the market. Using unified evaluation standards and algorithms, it ensures that the ratio schemes evaluated and selected each time are consistent and repeatable, avoiding decision fluctuations caused by human factors.

[0186] Step S3500, analyzing the production requirements of cement pole products corresponding to the current sub-order, extracting performance indicators, and establishing a performance indicator set; setting performance indicator thresholds, and dynamically adjusting the performance indicator thresholds according to the urgency of the current sub-order and the expected sales profit, and adaptively selecting the optimal ratio scheme from the optimal ratio scheme set;

[0187] Further, step S3500 includes:

[0188] Step S3510, analyzing the production requirements of cement pole products corresponding to the current sub-order, extracting performance indicators, and establishing a performance indicator set;

[0189] Performance indicator set KPI = {Q, R, S, E}, where Q is the raw material quality requirement, requiring the quality of each raw material to meet the specified standard; R is the ratio range, requiring the usage ratio of each raw material to be within the specified range; S is the strength grade requirement, requiring the product's compression, flexural, cracking and other strength indicators to meet the corresponding grade standards; E is the production efficiency requirement, requiring the production time, energy consumption, etc. to meet the specified upper limit;

[0190] Step S3520, for each performance indicator in the performance indicator set, comprehensively considering industry standards, customer requirements and enterprise quotas, setting a performance indicator threshold;

[0191] For the raw material quality requirement Q, set the raw material quality threshold Q0, requiring the quality of each raw material to reach Q0;

[0192] For the ratio range R, set the lower limit R of the amount of each raw material min And the upper limit of the dosage ratio R max ;

[0193] For strength level S, set the lower limit of strength index S min ;

[0194] For production efficiency E, set the upper limit of production time T max and energy consumption cap J max .

[0195] Step S3530, obtaining the urgency and expected sales profit of the current sub-order, and dynamically adjusting the performance indicator threshold according to the urgency and expected sales profit of the order;

[0196] Let the urgency of the current sub-order be U, U∈[0,1], the larger the value, the more urgent the order;

[0197] Assume that the expected sales profit is P, P∈[0,1], the larger the value, the higher the profit margin;

[0198] When U>U0, the performance indicator threshold is adjusted as follows:

[0199] Q0=Q0+ΔQ;

[0200] R min =R min -ΔR, R max =R max +ΔR;

[0201] S min =S min +ΔS;

[0202] T max =T max -ΔT,J max=J max -ΔJ;

[0203] in:

[0204] U0: is the preset urgency threshold;

[0205] ΔQ: When U>U0, the preset adjustment step of the raw material quality threshold;

[0206] ΔR: When U>U0, the preset adjustment step length of the lower limit and upper limit of the usage ratio;

[0207] ΔS: When U>U0, the preset adjustment step of the lower limit of the strength index;

[0208] ΔT: When U>U0, the adjustment step of the upper limit of production time;

[0209] ΔJ: When U>U0, the preset adjustment step of the energy consumption upper limit;

[0210] When P>P0, the performance indicator threshold is adjusted as follows:

[0211] Q0 = Q0 - ΔQ′;

[0212] R min =R min +ΔR′,R max =R max -ΔR′;

[0213] S min =S min -ΔS′;

[0214] T max =T max +ΔT′,J max =J max +ΔJ′;

[0215] in:

[0216] P0: is the preset sales profit threshold;

[0217] ΔQ': When P>P0, the preset adjustment step of the raw material quality threshold;

[0218] ΔR': When P>P0, the preset adjustment step length of the lower limit and upper limit of the dosage ratio;

[0219] ΔS': When P>P0, the preset adjustment step of the lower limit of the strength index;

[0220] ΔT': When P>P0, the adjustment step of the upper limit of production time;

[0221] ΔJ': When P>P0, the preset adjustment step of the energy consumption upper limit;

[0222] The step S3530 quantitatively characterizes the production requirements of the sub-orders by extracting performance indicators and setting performance indicator thresholds, and adaptively and dynamically adjusts the thresholds by introducing urgency and sales profit, so that the optimal matching plan can flexibly adapt to actual business needs.

[0223] Step S3540, traverse the optimal matching scheme set, determine whether each matching scheme meets the dynamically adjusted performance indicator threshold condition, and the scheme that meets the performance indicator threshold condition enters the alternative scheme pool. When the matching scheme does not meet the performance indicator threshold condition, determine whether the alternative scheme pool is empty; when the alternative scheme pool is not empty, enter step 3550; when the alternative scheme pool is empty, adjust the performance indicator threshold, and re-determine whether the matching scheme meets the adjusted performance indicator threshold condition;

[0224] When the pool of alternative solutions is empty, methods for adjusting the performance indicator threshold include:

[0225] Q0=Q0-ΔQ″;

[0226] R min =R min +ΔR″,R max =R max -ΔR″;

[0227] S min =S min -ΔS″;

[0228] T max =T max +ΔT″,J max =J max +ΔJ″;

[0229] in:

[0230] ΔQ”: When the alternative solution pool is empty, the preset adjustment step of the raw material quality threshold;

[0231] ΔR”: When the alternative solution pool is empty, the preset adjustment step size of the lower limit and upper limit of the usage ratio;

[0232] ΔS”: When the alternative solution pool is empty, the preset adjustment step of the lower limit of the strength index;

[0233] ΔT”: When the alternative solution pool is empty, the preset adjustment step of the production time upper limit;

[0234] ΔJ”: When the alternative solution pool is empty, the preset adjustment step of the energy consumption cap;

[0235] Step S3550, extracting the predicted material cost values ​​of the mix solutions in the pool of alternative solutions, and selecting the mix solution with the lowest cost as the optimal mix solution;

[0236] Step S3500 firstly extracts and analyzes the performance indicators comprehensively, so that the system can accurately evaluate the production requirements of each sub-order and ensure the scientificity and rationality of the matching scheme. Secondly, the performance indicator threshold is adjusted dynamically, so that the system can flexibly respond to the priorities and profit expectations of different orders, thereby improving resource utilization and production efficiency. Finally, by adaptively selecting the optimal matching scheme, the system can quickly find the optimal solution among multiple alternatives, significantly improving production efficiency and the timeliness of order delivery. Overall, this method effectively balances production requirements, urgency and profit expectations, ensures that each sub-order can be produced with the optimal matching scheme, and improves the company's production management level and market competitiveness.

[0237] Step S3600, generating a cement pole production formula for the current sub-order according to the optimal ratio scheme, adding it to the production formula set of the cement pole order, and updating the order progress information;

[0238] Step S3700, determine whether production formulas have been generated for all sub-orders. If so, output the complete production formula set for this cement pole order; if not, select the next sub-order and go to step S3100.

[0239] Step S3000 can customize and design multiple feasible proportioning schemes according to the specific needs of each sub-order, thereby improving the flexibility and adaptability of the scheme. By using the constructed cement pole proportioning optimization model, the material cost prediction, strength performance prediction and production efficiency prediction of each scheme in the proportioning scheme candidate set can be carried out, and the economy, performance and production efficiency of each scheme can be accurately evaluated. This process ensures that each sub-order minimizes production costs and maximizes production efficiency under the premise of meeting technical requirements. By screening out the optimal scheme in the proportioning scheme candidate set, the system can automatically select the proportioning scheme that best suits the current sub-order, reducing the error and workload of human decision-making. This not only improves the automation and efficiency of order processing, but also effectively improves the overall quality and economic benefits of cement pole production, thereby bringing significant competitive advantages to enterprises. Overall, step S3000 and its sub-steps optimize the proportioning process of cement poles through systematic and intelligent methods, ensure production accuracy, efficiency and cost control, and significantly improve the production management level and market responsiveness of enterprises.

[0240] Example 2

[0241] This embodiment provides an automatic proportioning and mixing system for cement pole raw materials based on order requirements on the basis of embodiment 1, such as Figure 3 As shown, including:

[0242] Order management module: used to receive cement pole orders, extract key order information and form order feature vectors, and verify and complete the order feature vectors.

[0243] Order splitting module: used to split the cement pole order into multiple sub-orders according to the order feature vector and the preset order splitting rule library to generate a sub-order set.

[0244] Historical data collection module: used to collect historical order demand data, historical production formula data, historical strength performance test data, historical material cost data and historical production efficiency data of cement poles, and clean, filter and extract features from these data to build a sample data set.

[0245] Ratio design module: Design a cement pole ratio optimization model based on historical production data. The cement pole ratio optimization model is a multi-objective optimization neural network model, which includes a shared backbone network and three independent sub-networks. The backbone network extracts high-dimensional features of order demand characteristics and formula parameter characteristics, and outputs them to three sub-networks. Each sub-network predicts material cost, strength performance and production efficiency respectively. The loss function of each sub-network together constitutes the loss function of the entire multi-objective optimization neural network model.

[0246] Scheme screening module: used to traverse the sub-order set, extract the key information of each sub-order, generate the ratio variables of the current sub-order according to the key information of the current sub-order; generate a candidate set of ratio schemes according to the ratio variables; obtain the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the candidate set of ratio schemes according to the key information of the current sub-order, the candidate set of ratio schemes and the cement pole ratio optimization model; according to the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the candidate set of ratio schemes, screen out the optimal ratio scheme in the candidate set of ratio schemes.

[0247] Example 3

[0248] This embodiment discloses an electronic device, including a memory, a central processing unit, and a computer program stored in the memory and executable on the central processing unit. When the central processing unit executes the computer program, the above-mentioned automatic proportioning and mixing method of cement pole raw materials based on order requirements is implemented.

[0249] Since the electronic device introduced in this embodiment is an electronic device used to implement the method for automatically proportioning and mixing raw materials for cement poles based on order requirements in the embodiment of this application, based on the method for automatically proportioning and mixing raw materials for cement poles based on order requirements introduced in the embodiment of this application, a person skilled in the art can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application will not be described in detail here. As long as a person skilled in the art implements the electronic device used by the method for automatically proportioning and mixing raw materials for cement poles based on order requirements in the embodiment of this application, it falls within the scope of protection of this application.

[0250] Example 4

[0251] The present embodiment discloses a computer-readable storage medium having a computer program stored thereon. When the computer program is executed, the above-mentioned method for automatically proportioning and mixing raw materials for cement poles based on order requirements is implemented.

[0252] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters, weights and thresholds in the formula are set by technicians in this field according to actual conditions.

[0253] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center through a wired network or a wireless network. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD) or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0254] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0255] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0256] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0257] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0258] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0259] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0260] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An automatic proportioning and mixing method for cement pole raw materials based on order requirements, characterized in that: include: Receive cement pole orders, extract key order information, and form order feature vectors based on the key order information; The cement pole order is split into multiple sub-orders according to the order feature vector and the preset order splitting rule library to generate a sub-order set; Collect historical production data of cement poles and build a cement pole ratio optimization model based on the historical production data; Traverse the sub-order set, extract the key information of each sub-order, generate the current sub-order ratio variable according to the key information of the current sub-order; generate the ratio scheme candidate set according to the ratio variable; obtain the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio scheme in the ratio scheme candidate set according to the key information of the current sub-order, the ratio scheme candidate set and the cement pole ratio optimization model; According to the material cost prediction value, strength property prediction value and production efficiency prediction value of each mix scheme in the mix scheme candidate set, the optimal mix scheme in the mix scheme candidate set is screened out; The cement pole ratio optimization model constructed according to historical production data includes: Collect historical production data of cement poles, wherein the historical production data includes historical order demand data, historical production formula data, historical strength performance test data, historical material cost data and historical production efficiency data; Clean, filter and extract features of historical production data to obtain order demand features, formula parameter features, strength performance features, material cost features and production efficiency features, and construct a sample data set based on order demand features, formula parameter features, strength performance features, material cost features and production efficiency features; The cement pole ratio optimization model is a multi-objective optimization neural network model, which includes a backbone network and sub-networks. The sub-networks include a material cost sub-network, a strength performance sub-network and a production efficiency sub-network, which are used to predict material cost, strength performance and production efficiency respectively; the backbone network receives order demand features and formula parameter features, extracts high-dimensional features of order demand features and formula parameter features through a fully connected layer, and generates a high-dimensional feature vector; each sub-network receives the high-dimensional feature vector output by the backbone network, and maps it to the corresponding material cost prediction value, strength performance prediction value and production efficiency prediction value through a fully connected layer; The loss functions of the material cost sub-network, the strength performance sub-network and the production efficiency sub-network are designed respectively, and the loss function of the entire multi-objective optimization neural network model is designed according to the loss functions of the material cost sub-network, the strength performance sub-network and the production efficiency sub-network; The generation of the matching variables of the current sub-order according to the key information of the current sub-order includes: According to the key information of the current sub-order, the material selection and process matching of cement poles are carried out to obtain the raw material types and production processes; Based on the key information, raw material types and production processes of the current sub-order, the usage percentage range of each raw material for cement poles is determined, and the ratio variables of the current sub-order are generated.

2. The method for automatically proportioning and mixing raw materials of cement poles based on order requirements according to claim 1 is characterized in that: The method comprises: The order splitting rule base is expressed in the form of IF-THEN production. The IF part is the order feature vector, and the THEN part is the splitting rule corresponding to the order feature vector. Automatically receive cement pole orders through the API interface of the order management system, use natural language processing technology to perform semantic analysis on the cement pole order text, and extract key order information; Match the order feature vector with the IF part in the order splitting rule library to find all splitting rules applicable to the current order; If multiple order splitting rules are matched, sort and filter them according to their priority and confidence, and select the optimal order splitting rule; Call the optimal order splitting rule to split the cement pole order into multiple sub-orders; Perform a legality check on the sub-orders to determine whether the sub-orders meet the requirements of the cement pole order. If they do not meet the requirements, adjust the order splitting rules and split the cement pole order again according to the adjusted order splitting rules. If they meet the requirements, continue to determine whether all sub-orders have undergone legality checks. If so, summarize all sub-orders to form a sub-order set; deduplicate and merge the sub-order set, and group and sort them according to the key order information; if not, perform a legality check on the next sub-order.

3. The method for automatically proportioning and mixing raw materials of cement poles based on order requirements according to claim 1 is characterized in that: The method comprises: According to the material cost prediction value, strength performance prediction value and production efficiency prediction value of the mix ratio scheme, an evaluation point in the three-dimensional space of cost-strength-efficiency is constructed, and the comprehensive evaluation score of each mix ratio scheme is obtained according to the evaluation point; According to the comprehensive evaluation scores of the matching schemes, the matching schemes in the matching scheme candidate set are sorted to obtain a matching scheme sequence from best to worst, which constitutes the optimal matching scheme set for the current sub-order; Analyze the production requirements of cement pole products corresponding to the current sub-order, and adaptively select the optimal ratio scheme from the set of optimal ratio schemes; Generate the cement pole production formula for the current sub-order based on the optimal ratio plan, add it to the production formula set of the cement pole order, and update the order progress information; generate cement pole production formulas for all sub-orders, and output the complete production formula set of this cement pole order.

4. The method for automatically proportioning and mixing raw materials of cement poles based on order requirements according to claim 1 is characterized in that: Determining the usage percentage range of each raw material of cement poles and generating ratio variables includes: Estimate the total volume based on the number, length and cross-sectional dimensions of the poles in the current sub-order, and convert the total concrete usage; estimate the cement usage range for each cubic meter of concrete based on the strength grade and production process; estimate the usage range of other materials by referring to the concrete mix design regulations; divide the usage range of each material by the total concrete usage to obtain the usage percentage range of each raw material for cement poles, and generate the mix ratio variables of the current sub-order based on the usage percentage range of each raw material for cement poles.

5. The method for automatically proportioning and mixing raw materials of cement poles based on order requirements according to claim 3 is characterized in that: An evaluation point in the three-dimensional space of construction cost-intensity-efficiency, and a comprehensive evaluation score of each ratio scheme obtained according to the evaluation point includes: Divide the material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix ratio scheme in the mix ratio scheme candidate set by the maximum value of the corresponding prediction value to perform dimensionless processing on the indicators; The dimensionless material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix ratio scheme are used as coordinates to mark the corresponding evaluation points in the three-dimensional space, and each evaluation point corresponds to a mix ratio scheme; In the three-dimensional space composed of the material cost prediction value, the strength performance prediction value, and the production efficiency prediction value, the coordinates of the ideal optimal value point and the worst value point are calculated, where the optimal value of the material cost prediction value takes the minimum value, and the worst value takes the maximum value; the optimal value of the strength performance prediction value and the production efficiency prediction value takes the maximum value, and the worst value takes the minimum value; The coordinates of the optimal value point and the coordinates of the worst value point constitute the ideal point and the negative ideal point respectively; Calculate the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point of each matching scheme; The comprehensive evaluation score of the matching scheme is calculated based on the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point.

6. The method for automatically proportioning and mixing raw materials of cement poles based on order requirements according to claim 5 is characterized in that: The method for calculating the comprehensive evaluation score of the matching scheme according to the Euclidean distance from the evaluation point to the ideal point and the Euclidean distance to the negative ideal point includes: Among them, S i is the comprehensive evaluation score of the i-th matching scheme, d i * is the Euclidean distance from the evaluation point to the ideal point of the i-th matching scheme, d i - is the Euclidean distance from the evaluation point to the negative ideal point of the i-th matching scheme, and γ is an adjustment parameter used to control the sensitivity of the score.

7. The method for automatically proportioning and mixing raw materials of cement poles based on order requirements according to claim 3 is characterized in that: The step of adaptively selecting the optimal matching scheme from the set of optimal matching schemes includes: Analyze the production requirements of cement pole products corresponding to the current sub-order, extract performance indicators, and establish a performance indicator set, including: performance indicator set KPI = {Q, R, S, E}, where Q is the raw material quality requirement, R is the ratio range, S is the strength grade requirement, and E is the production efficiency requirement; For each performance indicator in the performance indicator set, the performance indicator threshold is set by comprehensively considering industry standards, customer requirements and enterprise quotas; For the raw material quality requirement Q, set the raw material quality threshold Q0; For the ratio range R, set the lower limit R of the amount of each raw material min And the upper limit of the dosage ratio R max ; For strength level S, set the lower limit of strength index S min ; For production efficiency E, set the upper limit of production time T max and energy consumption cap J max ; Obtain the urgency and expected sales profit of the current sub-order, and dynamically adjust the performance indicator threshold according to the urgency and expected sales profit of the order; Traverse the set of optimal ratio schemes, determine whether each ratio scheme meets the dynamically adjusted performance indicator threshold conditions, and the schemes that meet the performance indicator threshold conditions enter the alternative scheme pool. When the ratio scheme does not meet the performance indicator threshold conditions, determine whether the alternative scheme pool is empty; when the alternative scheme pool is empty, adjust the performance indicator threshold, and re-determine whether the ratio scheme meets the adjusted performance indicator threshold conditions; when the alternative scheme pool is not empty, extract the material cost forecast value for the ratio scheme in the alternative scheme pool, and select the ratio scheme with the lowest cost as the optimal ratio scheme.

8. The method for automatically proportioning and mixing raw materials of cement poles based on order requirements according to claim 7 is characterized in that: The dynamically adjusting the performance indicator threshold according to the urgency of the order and the expected sales profit includes: Let the urgency of the current sub-order be U, U∈[0,1]; let the expected sales profit be P, P∈[0,1]; When U>U0, the performance indicator threshold is adjusted as follows: Q0=Q0+ΔQ;R min =R min -ΔR,R max =R max +ΔR; S min =S min +ΔS;T max =T max -ΔT,J max =J max -ΔJ; Among them, U0 is the preset urgency threshold; ΔQ is the preset adjustment step of the raw material quality threshold when U>U0; ΔR is the preset adjustment step of the lower limit and upper limit of the usage ratio when U>U0; ΔS is the preset adjustment step of the lower limit of the strength index when U>U0; ΔT is the preset adjustment step of the upper limit of the production time when U>U0; ΔJ is the preset adjustment step of the upper limit of energy consumption when U>U0; When P>P0, the performance indicator threshold is adjusted as follows: Q0=Q0-ΔQ′;R min =R min +ΔR′,R max =R max -ΔR′; S min =S min -ΔS′;T max =T max +ΔT′,J max =J max +ΔJ′; Among them, P0 is the preset sales profit threshold; ΔQ' is the preset adjustment step of the raw material quality threshold when P>P0; ΔR' is the preset adjustment step of the lower limit and upper limit of the usage ratio when P>P0; ΔS' is the preset adjustment step of the lower limit of the strength index when P>P0; ΔT' is the preset adjustment step of the upper limit of the production time when P>P0; ΔJ' is the preset adjustment step of the upper limit of energy consumption when P>P0.

9. An automatic proportioning and mixing system for cement pole raw materials based on order requirements, which is used to implement the automatic proportioning and mixing method for cement pole raw materials based on order requirements as claimed in any one of claims 1 to 8, characterized in that: include: Order management module: used to receive cement pole orders, extract key order information and form order feature vectors, and verify and complete the order feature vectors; Order splitting module: used to split the cement pole order into multiple sub-orders according to the order feature vector and the preset order splitting rule library, and generate a sub-order set; Historical data collection module: used to collect historical order demand data, historical production formula data, historical strength performance test data, historical material cost data and historical production efficiency data of cement poles, and perform data cleaning, screening and feature extraction to build a sample data set; Ratio design module: designs a cement pole ratio optimization model based on historical production data. The cement pole ratio optimization model is a multi-objective optimization neural network model, which includes a shared backbone network and three independent sub-networks. The backbone network extracts high-dimensional features of order demand characteristics and formula parameter characteristics, and outputs them to three sub-networks. Each sub-network predicts material cost, strength performance and production efficiency respectively. The loss function of each sub-network together constitutes the loss function of the entire multi-objective optimization neural network model. Solution screening module: used to traverse the sub-order set, extract the key information of each sub-order, generate the ratio variable of the current sub-order according to the key information of the current sub-order; generate the ratio solution candidate set according to the ratio variable; obtain the material cost prediction value, strength performance prediction value and production efficiency prediction value of each ratio solution in the ratio solution candidate set according to the key information of the current sub-order, the ratio solution candidate set and the cement pole ratio optimization model; According to the material cost prediction value, strength performance prediction value and production efficiency prediction value of each mix scheme in the mix scheme candidate set, the optimal mix scheme in the mix scheme candidate set is screened out.

Images