Intelligent batching system for precoated sand

Abstract

The invention relates to the field of intelligent manufacturing and industrial automation, and discloses a precoated sand intelligent batching system, which comprises an order information intelligent acquisition and verification module for receiving and verifying order information of a customer; the dynamic production scheduling engine module is used for dynamically arranging a production plan according to the order information and the production state; the multi-mode sensing weighing control module is used for collecting components and weights of crude sand and waste sand in real time and carrying out data fusion and compensation processing; the intelligent waste sand regeneration formula generation module is used for generating a mixed formula of regenerated sand so as to realize recycling of the waste sand; and the real-time monitoring and self-adaptive optimization module is used for carrying out real-time monitoring and dynamic optimization on the whole system, and ensuring that all the modules work cooperatively and have good response capability. Through the combination of data acquisition and real-time analysis, the smoothness of information flow is ensured, the weight and component data obtained in real time can be quickly processed, and the execution of a production plan is more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing and industrial automation, in particular to a coated sand intelligent batching system. BACKGROUND

[0002] In current production systems, static monitoring methods are commonly used. This approach usually relies on regular data acquisition, resulting in outdated information updates. As a result, any sudden conditions in the production process often cannot be reflected in a timely manner, resulting in low efficiency.

[0003] Many modules of a system operate independently of each other. Although each module has a clear function, there is a lack of effective cooperation. Resource allocation often cannot achieve the best, resulting in waste of funds and time. This independent operation method obviously fails to realize the potential benefits of the overall system.

[0004] In addition, existing technologies often rely on empirical rules to adjust production parameters. This method lacks flexibility and is difficult to respond to rapidly changing market demands. Fixed processes make the response time in the production process longer, increasing costs.

[0005] The lack of feedback mechanisms is also a significant shortcoming. Most current systems lack real-time feedback capabilities, making it difficult to effectively correct production conditions. This results in a lack of adaptability when the system faces changes, making it difficult to maintain efficient operation.

[0006] Therefore, the deficiencies of existing technologies are obvious, both in terms of information lag and module coordination, and lack of flexibility. These problems collectively constrain the improvement of production efficiency, and an even more efficient and intelligent solution is urgently needed. SUMMARY

[0007] To address the deficiencies of existing technologies, the present application provides a coated sand intelligent batching system, which solves the problems of information lag, module independent operation and insufficient feedback in traditional production systems.

[0008] To achieve the above purpose, the present application realizes the following technical scheme: a coated sand intelligent batching system, comprising:

[0009] An order information intelligent acquisition and verification module for receiving and verifying customer order information;

[0010] A dynamic production scheduling engine module connected to the order information intelligent acquisition and verification module for dynamically arranging production plans according to order information and production status;

[0011] A multi-modal perception weighing control module connected to the dynamic production scheduling engine module for real-time acquisition of the composition and weight of raw sand and waste sand, and data fusion and compensation processing;

[0012] The waste sand intelligent regeneration formula generation module is connected with the multi-modal perception weighing control module, and is used for generating a mixed formula of regenerated sand, so as to realize the recycling of waste sand.

[0013] The real-time monitoring and adaptive optimization module is connected with all other modules, and is used for real-time monitoring and dynamic optimization of the whole system.

[0014] Further, the order information intelligent acquisition and verification module: this module is responsible for receiving and verifying customer order information. Through multiple input methods, it ensures that all order information is accurate and reliable, providing reliable basic data for subsequent production scheduling and planning.

[0015] Dynamic production scheduling engine module: this module is connected with the order information intelligent acquisition and verification module, and is responsible for dynamically arranging production plans according to order information and current production status. It optimizes production arrangement using Markov decision process, and automatically adjusts production strategy based on real-time working conditions through reward function to improve production efficiency and flexibility.

[0016] Multi-modal perception weighing control module: connected with the dynamic production scheduling engine module, this module real-time acquires composition and weight data of raw sand and waste sand, and performs data fusion and compensation processing. The hyperspectral imaging device of this module works together with the weighing sensor to ensure accurate tracking and management of raw materials during production.

[0017] Waste sand intelligent regeneration formula generation module: this module is connected with the multi-modal perception weighing control module, and is responsible for generating a mixed formula of regenerated sand according to real-time acquired composition and weight data. This process models waste sand composition through graph neural network to optimize the proportion of regenerated sand, ensuring effective utilization and regeneration of waste sand.

[0018] Real-time monitoring and adaptive optimization module: this module is connected with all other modules, and is responsible for real-time monitoring and dynamic optimization of the whole production system. It real-time monitors production speed, raw material usage, production status and other parameters, and generates optimization suggestions through data analysis to improve the overall performance and efficiency of the production system.

[0019] Preferably, the order information intelligent acquisition and verification module receives order data through multiple input methods, including electronic orders, scanned paper orders and manual input.

[0020] Further, through multiple input methods, order data is flexibly received, increasing the convenience and accuracy of information acquisition, ensuring that different forms of orders can be quickly processed, and improving the timeliness and effectiveness of production planning.

[0021] Preferably, the dynamic production scheduling engine module builds a Markov decision process to optimize the arrangement of production plans and automatically adjusts according to real-time working conditions through a reward function, and the specific reward function is calculated as follows:

[0022] R t = w1(t) r efficiency + w2(t) r energy + w3(t) r quality ;

[0023] In the formula, R t represents the comprehensive score or performance evaluation at time t, w1(t) is the weight related to efficiency, w2(t) is the weight related to energy consumption, w3(t) is the weight related to quality, r efficiency represents the score related to efficiency, r energy represents the score related to energy consumption, and r quality represents the score related to quality.

[0024] Further, the Markov decision process and the reward function are used to optimize the production plan, so that it has self-adaptive ability and can automatically adjust the production strategy according to real-time working conditions, and finally realizes the balance between efficiency, energy consumption and product quality.

[0025] Preferably, the multi-modal perception weighing control module includes a hyperspectral imaging device and a weighing sensor.

[0026] Further, by introducing the hyperspectral imaging device and the weighing sensor, the real-time monitoring capability of the composition and the weight is strengthened, and the data obtained is more comprehensive, so that the subsequent dynamic proportioning compensation is more accurate.

[0027] Preferably, the dynamic proportioning compensation calculation of the multi-modal perception weighing control module uses the following formula for adjustment:

[0028] W adjusted = W base + a AC + b AT + g ||F fused ||2;

[0029] In the formula, W adjusted is the adjusted weight, W base is the basic weight, AC is the composition deviation, AT is the temperature drift, ||F fused ||2 is the two-dimensional module of force, and a, b, g are adaptive compensation coefficients.

[0030] Specifically, the adjustment through the preset formula can effectively handle the composition deviation and temperature influence, and ensure that the best material proportioning is always maintained in actual production, thereby improving the consistency and quality of the product.

[0031] Preferably, the specific steps of the dynamic proportioning compensation calculation include: obtaining the current component deviation and temperature drift; calculating the adjusted weight according to the preset basic weight and the corresponding compensation coefficient; and applying the adjusted weight to the actual material proportioning.

[0032] Preferably, the compensation coefficient used in the dynamic proportioning compensation calculation is obtained by training a machine learning model based on historical data.

[0033] Further, the steps of dynamic proportioning compensation calculation are defined in detail, and by obtaining the component deviation and temperature drift, real-time adjustment of material proportioning is realized, ensuring that the reclaimed sand can meet the production requirements.

[0034] Preferably, the waste sand intelligent regeneration formula generation module models the components of waste sand through a graph neural network and uses physical constraints to optimize the proportioning of reclaimed sand, which ensures that the component proportions of reclaimed sand meet the process requirements.

[0035] Further, the proportioning of reclaimed sand is optimized using a graph neural network and physical constraints to ensure that the generated formula meets the process requirements, thereby achieving efficient utilization of waste sand and achieving the goals of environmental protection and resource recycling.

[0036] Preferably, the operation steps of the multi-modal perception weighing control module include: component analysis of raw sand and waste sand through a hyperspectral imaging device; real-time weight data of raw sand and waste sand are obtained through a weighing sensor; and the component and weight data are input into a dynamic proportioning compensation algorithm for proportioning adjustment.

[0037] Preferably, the real-time monitoring and adaptive optimization module is used to monitor various parameters in the production process in real time, including but not limited to production speed, raw material usage, and production status, and generate optimization suggestions through data analysis.

[0038] Further, through real-time monitoring and data analysis of production parameters, this module can timely discover and solve problems that may occur in production, provide optimization suggestions to promote the improvement of production efficiency, and ensure the efficiency and stability of the overall process.

[0039] The present application provides a coated sand intelligent batching system. It has the following advantages:

[0040] 1、The present application adopts a real-time monitoring and adaptive optimization technical solution, which realizes comprehensive monitoring and dynamic adjustment of the production system. In this way, the system can quickly respond to real-time changes and improve production efficiency. Compared with the static monitoring solution in the prior art, this method effectively solves the problem of slow production caused by information lag.

[0041] 2、The present application ensures smooth information flow by combining data collection and real-time analysis, and the weight and composition data obtained in real time can be quickly processed. This efficient data interaction avoids the delay in data updating of the past, making the execution of production plans more accurate.

[0042] 3、The present application introduces a dynamic optimization algorithm, which enables better cooperation between modules. By flexibly adjusting resource allocation and device priority, it ensures integrated operation. The existing system often operates independently, leading to resource waste and low efficiency.

[0043] 4、The present application also establishes a feedback mechanism to form a closed-loop control system, improving the system's self-adaptive ability. This design allows production to be continuously adjusted according to actual conditions, avoiding the limitations of fixed processes in the past, greatly enhancing the ability to respond to changes. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 The present application is a schematic diagram of the system framework;

[0045] Figure 2 The present application is a schematic diagram of the order information intelligent acquisition and verification module;

[0046] Figure 3 The present application is a schematic diagram of the dynamic production scheduling engine module;

[0047] Figure 4 The present application is a schematic diagram of the multi-modal sensing and weighing control module;

[0048] Figure 5 The present application is a schematic diagram of the waste sand intelligent regeneration formula generation module;

[0049] Figure 6 The present application is a schematic diagram of the real-time monitoring and adaptive optimization module. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the present application will be described in detail below with reference to the accompanying drawings of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0051] Please refer to the accompanying drawings of the present application Figure 1 -Appendix Figure 6 The present application provides an intelligent coating sand batching system, which comprises:

[0052] The order information intelligent acquisition and verification module is used to receive and verify the customer's order information;

[0053] Specifically, the system composition and structure:

[0054] The order information intelligent collection and verification module mainly consists of the following parts:

[0055] Data input interface: including electronic order receiving module, paper order scanning module and manual input module, which can flexibly handle different types of order input;

[0056] Data processing unit: to analyze and arrange the input data for verification and subsequent use;

[0057] Verification logic: including a series of rules and algorithms, used to verify the integrity and validity of the order;

[0058] Feedback mechanism: to feed back the processing results to the user, to ensure that the user can understand the order status in real time.

[0059] The above components are connected by logic to form a complete system, forming an effective data flow and processing flow, ensuring that each link can work together.

[0060] Detailed implementation of data input interface:

[0061] The data input interface has multiple information receiving methods to adapt to the diversified order needs of customers. The specific implementation includes:

[0062] Electronic order receiving module: can interact with customers' e-commerce platform or ERP system through API, data is received in JSON or XML format, and related information is extracted using parsing algorithm;

[0063] Paper order scanning module: using optical character recognition (OCR) technology, the paper order is converted into digital information through scanner, and the image data generated after scanning is parsed into recognizable text information by OCR algorithm;

[0064] Manual input module: provides a user-friendly input interface, users can manually input order information, the interface includes input box, drop-down menu and date selector, etc. Control to improve the accuracy and convenience of input;

[0065] All data is output to the data processing unit in a unified format after passing through the corresponding module.

[0066] Functions of data processing unit:

[0067] In the data processing unit, first of all, the input data is formatted, including removing unnecessary white space, unifying data format, so as to ensure data consistency. The specific steps are as follows:

[0068] Format check: using rule engine to check the format of each order field, such as customer name should be a string, quantity should be a positive integer, etc.

[0069] Data integrity check: ensure that all necessary fields in the order have been filled, such as customer name, product number, order quantity, delivery time, etc. Missing important fields will be marked as "invalid" and output corresponding error information.

[0070] After all format checks pass, the data will be transmitted to the verification logic module.

[0071] In the verification logic module, first set multiple verification rules and implement logical judgment for each rule. The key core step is variable validity judgment, which is processed using the following formula:

[0072] V i =f(D i );

[0073] In the formula, V i represents the validity status of the i-th order parameter, and D i is the actual value of the parameter. The specific algorithm process of function f is defined as follows:

[0074] Condition judgment:

[0075] Order quantity should be greater than zero, that is, if D i = quantity, then D i >0, V i is true.

[0076] Delivery time should be greater than or equal to the current system time, let the current time be T now , then: if D i = delivery time, the judgment condition is: D i ≥T now , at this time, if it is satisfied, V i is true.

[0077] Through the above condition judgment, if all parameters are valid, the output result is "pass the verification", and the data is transmitted to the dynamic production scheduling engine module. But if any parameter is invalid, it will be marked as "verification failed", and the corresponding error information will be stored for user to check.

[0078] Data flow and logical connection:

[0079] The logical connection of the entire module makes the data flow clear in each link. The input data enters the data processing unit through the data input interface, then enters the verification logic module, and finally outputs to the dynamic production scheduling engine module. The processing result will be updated to the user interface in real time through the feedback mechanism, ensuring that the user can obtain the order processing progress in time.

[0080] Implementation of feedback mechanism:

[0081] The feedback mechanism is designed to inform users of the processing status in real time, including the following:

[0082] Successful reception notification: When the user successfully submits an order, the system will inform the user of the reception status through a pop-up window or a prompt message;

[0083] Verification result notification: If the order passes the verification, it will feedback "order processing success"; if the verification fails, it will list the reasons in detail.

[0084] Historical data storage and management:

[0085] All successful order data and verification records are stored in the system database, using SQL database or NoSQL database storage solution to ensure the persistence and traceability of order data. This design not only helps subsequent data analysis, but also provides support for the decision-making of the dynamic scheduling engine.

[0086] In summary, this embodiment describes the internal logical relationship, structure composition and function implementation of the order information intelligent collection and verification module. Through efficient data processing and rigorous verification logic, this module ensures the accuracy and integrity of order information, laying a solid foundation for the operation of the entire platform. The implementation of this technical solution can effectively improve the order processing efficiency, reduce production problems caused by information errors, and adapt to diverse customer needs.

[0087] Dynamic production scheduling engine module, connected with order information intelligent collection and verification module, used for dynamically arranging production plan according to order information and production status;

[0088] Specifically, the design of dynamic production scheduling engine module consists of three main parts: data reception, decision calculation and production scheduling implementation. The module receives verified order data through standard interface, which contains important fields such as product type, quantity, delivery time, etc. These data are transmitted in a unified standard format (such as JSON or XML), ensuring the integrity and consistency of information.

[0089] After receiving order data, the dynamic production scheduling engine module first assesses the current production status. This assessment includes analyzing the production line's available resources, current production load, incoming orders, and production capacity. This status information comes from the real-time monitoring and adaptive optimization module, which uses sensors and data acquisition systems to obtain real-time information, ensuring the scheduling engine has a comprehensive understanding of the production environment.

[0090] After the state evaluation is completed, the reward function is automatically adjusted according to the real-time working conditions. The specific reward function calculation formula is:

[0091] R t =w1(t)·r efficiency +w2(t)·r energy +w3(t)·r quality ;

[0092] In the formula,

[0093] R t represents the comprehensive score or performance evaluation at time t, w1(t) is the weight related to efficiency, w2(t) is the weight related to energy consumption, w3(t) is the weight related to quality, and r efficiency represents the efficiency-related score, r energy represents the score related to energy expenditure, r quality Represents a quality-related score.

[0094] The specific implementation process includes the following steps: First, the training model uses historical data to generate initial weight coefficients. This process can apply machine learning algorithms such as linear regression and neural networks, and optimize the weight parameter settings by analyzing the success and failure cases of historical orders.

[0095] The scheduling engine then generates a production plan based on the real-time data and evaluation results. If the current plan cannot meet all order requirements, an automated rescheduling process is initiated. This process assesses the flexibility of each process and automatically adjusts the process sequence, resource allocation, and even line switching to ensure delivery deadlines are met.

[0096] For example, if a piece of production equipment fails, the system will readjust the production plan and production sequence in real time, giving priority to unaffected orders to reduce the impact on the overall production progress.

[0097] Implementing a monitoring loop at the core of the scheduling engine is crucial. This monitoring module provides real-time feedback on production progress and status to the scheduling engine. If the monitoring system detects that production speed at a particular stage is falling below the target, the scheduling engine adjusts subsequent plans based on this data to ensure all orders are completed on time.

[0098] In addition, the dynamic production scheduling engine module also contains a decision support system, providing data-based visual analysis, presenting information in the form of charts and dashboards on the user interface, enhancing the visibility of decision-making.

[0099] All scheduling decisions and status information will be stored in the database for subsequent analysis and improvement. This historical data can help the scheduling engine continue to optimize its decision model in future scheduling processes, enhancing the intelligence level of the system.

[0100] In summary, the dynamic production scheduling engine module integrates order information and real-time production status, uses mathematical models and machine learning algorithms for intelligent scheduling. This module not only realizes dynamic scheduling through accurate evaluation models, but also continuously optimizes production plans through real-time monitoring and historical data feedback. This implementation provides a solid foundation for the efficient operation of the entire coated sand intelligent batching system, ensuring that the system can respond quickly when facing changing market demands.

[0101] The multi-modal perception weighing control module is connected to the dynamic production scheduling engine module for real-time acquisition of the composition and weight of raw sand and waste sand, and performs data fusion and compensation processing.

[0102] Specifically, the multi-modal perception weighing control module is composed of several key components, including a hyperspectral imaging system, a weighing sensor, a data processing unit, and a fusion output module. The hyperspectral imaging system is used to acquire real-time composition information of raw sand and waste sand, the weighing sensor monitors the weight of the material in real time, the data processing unit processes and fuses the collected composition and weight data, and finally transmits the results to the dynamic production scheduling engine module.

[0103] In the implementation process, first, the hyperspectral imaging system scans the input raw sand and waste sand. The system emits light of different wavelengths and receives reflected light to analyze the spectral characteristics of each component. This process generates composition data, called C j , which is defined as:

[0104]

[0105] In the formula, C j represents the mass fraction of the jth component, m j is the mass of the component, M is the total mass of the sample, and n is the number of component types.

[0106] Next, the weighing sensor measures the total weight of raw sand and waste sand in real time, with the output value being W, in kilograms (kg). This data should be regularly calibrated to ensure the accuracy of the measurement. The data processing unit fuses the composition data and weight data, and uses the following formula to adjust and ensure accurate true weight:

[0107] W adjusted = W base + α·ΔC + β·ΔT + γ·||F fused ||2;

[0108] where W adjusted is the adjusted weight, W base is the base weight, ΔC is the composition deviation, ΔT is the temperature drift, ||F fused ||2 is the two-dimensional modulus of force, and α, β, γ are adaptive compensation coefficients.

[0109] During the adjustment process, first, the weighing results are baseline adjusted to ensure the accuracy of the measured data. For example, by regularly checking the weighing sensor, any drift or error is eliminated. Then, the composition deviation and temperature offset are calculated, which will significantly affect the accuracy of the weight measurement. Through statistical methods and control chart analysis, the values of α, β, and γ are dynamically adjusted.

[0110] The fused data will be organized as structured data and transmitted to the dynamic production scheduling engine module through standardized protocols such as MQTT or RESTful API. Based on these data, the engine performs more accurate production planning optimization, such as scheduling resources, adjusting production processes, or balancing the load of different production lines to ensure the continuity and efficiency of production.

[0111] This module also has real-time monitoring function, all ingredients, weights and their adjustment data will be stored in the database in real time, ensuring the traceability of data. Historical data can be used for further analysis, through machine learning algorithm to optimize the future measurement and adjustment model, improve the intelligent management level of the system.

[0112] For example, if the true proportion of a certain ingredient is found to be significantly different from the predicted data after each measurement, the system will record this information and train the prediction model to optimize the parameter settings. In addition, the system can introduce environmental monitoring data to adjust the impact of temperature and humidity on the weighing results in real time.

[0113] In summary, the multi-modal perception weighing control module provides key data support for the dynamic production scheduling engine module through efficient ingredient analysis and weight measurement. Through this implementation, the system has high accuracy, flexibility and intelligent characteristics, not only improves production efficiency, but also ensures the accuracy of material configuration, meets the needs of modern production. The design and implementation of this module fully demonstrates the technological innovation and application value, providing a solid foundation for the intelligentization of the production site.

[0114] The waste sand intelligent regeneration formula generation module is connected with the multi-modal perception weighing control module, and is used for generating a mixed formula of regenerated sand, so as to realize the recycling of waste sand.

[0115] Specifically, the waste sand intelligent regeneration formula generation module includes a component analysis unit, a formula optimization algorithm module, a formula generation output module, and a data interface. These components are logically connected and cooperate with each other, enabling the entire system to operate efficiently.

[0116] The component analysis unit receives component data C j and weight data W adjusted from the multi-modal perception weighing control module.

[0117] The weight data W adjusted represents the adjusted total weight of the waste sand. This data is influenced by the multi-modal perception weighing control module to ensure its accuracy and reliability. The formula generation module establishes a mathematical model based on the target performance set by the user, such as strength, fluidity, plasticity, etc. Set the target performance indicator P target , use the following formula to predict performance:

[0118]

[0119] In the formula, P target represents the target performance indicator, represents the sum of all terms from j = 1 to n, w j is the contribution weight of the jth component to the indicator, C j is the mass fraction of the component, and n is the number of component types.

[0120] The formula optimization algorithm uses genetic algorithms, particle swarm optimization, or other intelligent optimization techniques for iterative adjustment. The optimization process mainly follows the following steps:

[0121] Initialization: Generate an initial population and randomly assign component proportions C j to meet the constraint conditions;

[0122] Fitness evaluation: Evaluate the fitness of each formula by calculating the performance indicator P of the current solution. This process can be achieved through the following formula:

[0123] Fitness = f(P);

[0124] In the formula, Fitness represents fitness, f(P) represents how fitness is calculated based on parameter P, and P represents a parameter related to fitness.

[0125] Selection operation: Select good formulas based on fitness values to generate a basis for subsequent crossover and mutation operations;

[0126] Crossover and mutation: Generate new formulas through crossover and mutation operations, and iteratively update C j ;

[0127] Termination condition: end the optimization when the preset termination condition is met (such as the fitness reaches the target or the upper limit of the number of iterations).

[0128] The final formula generated by the above optimization algorithm will be formatted and transmitted to the formula generation output module, which will organize the generated formula into a standardized data format for subsequent application in the production process.

[0129] In addition, the module is connected with the multi-modal perception weighing control module through a standard data interface, ensuring smooth real-time data flow. This connection relationship enables timely updating of component and weight data, reflecting the actual state of waste sand and enhancing the response capability of the entire system.

[0130] During implementation, the entire module will record the parameters of each formula generation and store this information in a database. Through analysis of historical data, the optimization algorithm can be effectively supported, enabling feedback adjustment.

[0131] To enhance the economic efficiency and sustainability of waste sand recycling, the system will integrate various cost control indicators. The cost function is set as follows:

[0132]

[0133] where Cost is the total cost, represents the summation of all terms from j = 1 to n, c j is the unit cost of the jth component.

[0134] Through this implementation, the waste sand intelligent regeneration formula generation module effectively combines component analysis and optimization algorithms to achieve efficient waste sand regeneration formula generation. This module fully demonstrates innovation and practicality, providing a scientific solution for sustainable use of waste sand and supporting the intelligent needs of modern production.

[0135] Real-time monitoring and adaptive optimization module, connected with all other modules, for real-time monitoring and dynamic optimization of the entire system;

[0136] Specifically, the real-time monitoring and adaptive optimization module includes a data acquisition unit, a real-time analysis unit, a decision support unit, and a feedback control unit. These components are logically connected to form an efficient monitoring and optimization system.

[0137] The data acquisition unit is responsible for obtaining real-time data from all relevant modules such as the multi-modal perception weighing control module, the waste sand intelligent regeneration formula generation module, and the dynamic production scheduling engine module. The acquired data includes the following categories:

[0138] Weight data W current : real-time measurement of the weight of waste sand and raw sand;

[0139] Component data C j : mass fraction of each component in waste sand;

[0140] Production state S: current working state of production line, such as running, standby or failure;

[0141] Scheduling information D: current production scheduling task, including execution state and expected time of each process.

[0142] The real-time analysis unit processes and analyzes these data. Multiple monitoring indicators M k are set to evaluate the working state of different modules, defined as follows:

[0143] M k = f(R k ,T k ,S k ) (k = 1, 2, …, n);

[0144] In the formula, M k represents the kth monitoring indicator, quantifying module performance, f(R k ,T k ,S k ) represents a function, calculating input parameters used when M k , R k represents an indicator of resource utilization efficiency, T k represents an indicator of time response, S k represents an indicator of system health state, and k = 1, 2, …, n represents the index k, ranging from 1 to n.

[0145] The real-time analysis unit will integrate all indicators to form system performance evaluation through weighting and standardization processing:

[0146]

[0147] In the formula, P system represents the overall performance value of the system, represents the sum of all terms from k = 1 to n, and w k represents the weight of the kth indicator.

[0148] The decision support unit adjusts the running state of each module based on real-time analysis results using dynamic optimization algorithms. Adopting feedback control mechanism in control theory, the goal is to optimize the overall performance of the system. The optimization process is as follows:

[0149] Objective setting: clearly define optimization goals, including reducing energy consumption, improving productivity, shortening reaction time, etc.

[0150] State evaluation: based on monitoring indicators M kAnalyzing the current state of each module;

[0151] Enhanced feedback: Adjusting control parameters α, β, γ for feedback control of modules, for example:

[0152] α represents the coefficient of adjusting the response sensitivity of a module;

[0153] β represents the setting of resource allocation priority;

[0154] γ represents the flexibility of scheduling optimization;

[0155] Dynamic strategy generation: Using reinforcement learning algorithm to generate optimal strategy, ensuring the adaptive ability of the system and improving the collaborative effect between different modules.

[0156] Finally, the optimized control signal will be fed back to each relevant module through standardized protocols (such as MQTT or RESTful API), ensuring the efficient response of the system. This process realizes dynamic adjustment through timely feedback, so that each module can adapt to the changes in current production demand.

[0157] Throughout the implementation process, the connection relationship between the real-time monitoring and adaptive optimization module and other modules is very clear. The data acquisition unit obtains input from each module, the real-time analysis unit is responsible for processing and evaluating data, and the decision support unit performs dynamic optimization based on the analysis results.

[0158] The output of this module will be recorded and stored in the database, forming historical data. These data not only help subsequent analysis, but also will be used for continuous improvement of optimization algorithms, forming an efficient feedback mechanism.

[0159] In summary, the real-time monitoring and adaptive optimization module ensures the collaborative working ability and good response speed of the entire system through efficient data acquisition, real-time analysis and dynamic feedback. This implementation shows the innovation and effectiveness of the system, providing strong support for meeting the needs of modern production. Reasonable module connection and information flow ensure efficient use of resources and promote the continuous development of intelligent production.

[0160] Although embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and alterations can be made without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. An intelligent batching system for coated sand, characterized in that: include: Order information intelligent collection and verification module, used to receive and verify customer order information; Dynamic production scheduling engine module, connected to the order information intelligent collection and verification module, is used to dynamically arrange production plans based on order information and production status; The multimodal sensing weighing control module is connected to the dynamic production scheduling engine module to collect the composition and weight of raw sand and waste sand in real time, and perform data fusion and compensation processing; The waste sand intelligent regeneration formula generation module is connected to the multimodal sensing weighing control module to generate a mixed formula for regenerated sand, thereby realizing the reuse of waste sand; The real-time monitoring and adaptive optimization module is connected with all other modules to perform real-time monitoring and dynamic optimization of the entire system.

2. The intelligent batching system for coated sand according to claim 1, characterized in that: The order information intelligent collection and verification module receives order data through multiple input methods, including electronic orders, scanning of paper orders, and manual input.

3. The intelligent batching system for coated sand according to claim 1, characterized in that: The dynamic production scheduling engine module constructs a Markov decision process to optimize the production plan and automatically adjusts it according to the real-time working conditions through the reward function. The specific reward function calculation formula is: R t =w1(t)·r efficiency +w2(t)·r energy +w3(t)·r quality ; In the formula, R t represents the comprehensive score or performance evaluation at time t, w1(t) is the weight related to efficiency, w2(t) is the weight related to energy consumption, w3(t) is the weight related to quality, and r efficiency represents the efficiency-related score, r energy represents the score related to energy expenditure, r quality Represents a quality-related score.

4. The intelligent batching system for coated sand according to claim 1, characterized in that: The multimodal perception weighing control module includes a hyperspectral imaging device and a weighing sensor.

5. The intelligent batching system for coated sand according to claim 1, characterized in that: The dynamic ratio compensation calculation of the multimodal sensing weighing control module is adjusted using the following formula: W adjusted =W base +α·ΔC+β·ΔT+γ·||F fused ||2; In the formula, W adjusted is the adjusted weight, W base is the basic weight, ΔC is the composition deviation, ΔT is the temperature drift, ||F fused ||2 is the two-dimensional modulus of the force, and α, β, and γ are adaptive compensation coefficients.

6. The intelligent batching system for coated sand according to claim 1, characterized in that: The specific steps of the dynamic ratio compensation calculation include: obtaining the current component deviation and temperature drift; calculating the adjusted weight according to the preset basic weight and the corresponding compensation coefficient; and applying the adjusted weight to the actual material ratio.

7. The intelligent batching system for coated sand according to claim 1, characterized in that: The compensation coefficient used in the dynamic proportioning compensation calculation is obtained by training a machine learning model based on historical data.

8. The intelligent batching system for coated sand according to claim 1, characterized in that: The waste sand intelligent regeneration formula generation module models the composition of waste sand through a graph neural network and adopts physical constraints to optimize the ratio of regenerated sand. The physical constraints ensure that the composition ratio of the regenerated sand meets the process requirements.

9. The intelligent batching system for coated sand according to claim 1, characterized in that: The operating steps of the multimodal sensing weighing control module include: performing composition analysis on raw sand and waste sand through hyperspectral imaging equipment; then obtaining real-time weight data of raw sand and waste sand through weighing sensors; and inputting the composition and weight data into a dynamic ratio compensation algorithm to adjust the ratio.

10. The intelligent batching system for coated sand according to claim 1, characterized in that: The real-time monitoring and adaptive optimization module is used to monitor various parameters of the production process in real time, including but not limited to production speed, raw material usage and production status, and generate optimization suggestions through data analysis.

Images