Search

CN-121981445-A - Intelligent parameter recommendation system and method suitable for automatic line adjustment procedure of oil injector

CN121981445ACN 121981445 ACN121981445 ACN 121981445ACN-121981445-A

Abstract

An intelligent parameter recommendation system and method suitable for an automatic line adjustment procedure of an oil sprayer relates to the field of intelligent control of industrial production and comprises a data acquisition module, a data processing module, a model training module, an intelligent Agent module and a parameter execution module, wherein the data processing module is used for carrying out standardized processing on multi-source data acquired by the data acquisition module, extracting a propulsion parameter optimization key factor to generate a sample pair and carrying out structural storage, the model training module is used for selecting an AI model, carrying out incremental training based on the sample pair, learning a working condition and propulsion mapping relation and parameter adjustment experience, the intelligent Agent module is used for acquiring all-working condition data, calling the AI model to generate candidate parameter tables in batches, carrying out all-flow simulation evaluation and screening on the candidate parameter tables, writing the candidate parameter tables into a database by the parameter execution module, triggering a control system to download and synchronizing the candidate parameter tables to equipment, and enabling the equipment to be called autonomously according to real-time working conditions. The invention greatly improves the parameter adjustment accuracy and the production line adaptation efficiency, reduces the parameter adjustment workload of engineers, strengthens the operation stability and the product quality, and adapts to the industrial production requirements.

Inventors

  • MO TAO
  • ZENG CHUNLIAN
  • SHI DINGHUA

Assignees

  • 柳州源创电喷技术有限公司

Dates

Publication Date
20260505
Application Date
20251230

Claims (10)

  1. 1. The intelligent parameter recommendation system suitable for the automatic line adjustment procedure of the oil sprayer is characterized by comprising a data acquisition module, a data processing module, a model training module, an intelligent Agent module and a parameter execution module: The data acquisition module is used for docking the data acquisition system of the automatic line equipment of the oil sprayer so as to acquire real-time operation data of the equipment and historical production process data, and the acquired multi-source data are structured and synchronized to the data processing module; the data processing module is used for carrying out standardized processing on the multi-source data, extracting key influence factors related to propulsion parameter optimization, generating an input characteristic-optimal propulsion parameter sample pair, and simultaneously transmitting the standardized data and the standardized samples to the server storage unit for storage; The model training module is in data interaction with the server storage unit and is used for calling the standardized data and sample pairs from the server storage unit, selecting an AI model suitable for an industrial scene, performing incremental training and iterative optimization, enabling a model deep learning engineer to tune experience, equipment operation rules and mapping relation between parameters and process effects, and finally generating an AI parameter recommendation model with industrial scene parameter recommendation capability; the intelligent Agent module is used for carrying out data interaction with the model training module and the server storage unit and serving as an interaction and decision-making execution carrier of the AI parameter recommendation model, and is used for calling the standardized historical production process data from the server storage unit, calling the trained AI parameter recommendation model for reasoning, outputting the optimal propulsion parameters and the parameter adaptation degree score, and generating a standardized optimal propulsion parameter table; and the parameter execution module is in data interaction with the intelligent Agent module and the automatic line control system of the oil injector, and is used for writing an optimal propulsion parameter table into a database parameter table, triggering the automatic line control system of the oil injector to download the optimal propulsion parameter corresponding to the working condition from the database parameter table, and synchronizing the optimal propulsion parameter to the parameter configuration module of the equipment to execute the operation.
  2. 2. The intelligent parameter recommendation system suitable for the automatic line adjustment process of the fuel injector according to claim 1, further comprising a feedback optimization module which is respectively in communication connection with the parameter execution module, the automatic line equipment data acquisition system of the fuel injector and the model training module and is used for acquiring adjusted production data, quantitatively calculating adjustment effects and feeding back relevant data to the model training module to realize model iterative optimization.
  3. 3. The intelligent parameter recommendation system suitable for the automatic line adjustment procedure of the fuel injector according to claim 2, wherein the data processing module comprises a standardized processing unit and a feature extraction unit, the feature extraction unit is used for extracting key factors such as target flow, propulsion, qualification rate and the like, the standardized processing unit generates a sample pair of input feature-optimal propulsion parameters according to the unified dimension of the fuel injector parameter standard, and the standardized processed data and the sample pair are stored in a server storage unit, so that the full link traceability of the data is ensured.
  4. 4. The intelligent parameter recommendation system suitable for the automatic line adjustment procedure of the oil sprayer according to claim 3, wherein the model training module comprises a model selection unit, a prompt word optimization unit and an increment training unit, wherein the model selection unit is used for selecting an AI model suitable for an automatic line scene of the oil sprayer, the prompt word of the prompt word optimization unit comprises scene description, input parameters, current working condition parameters and output format specifications, the 'optimal propulsion quantity + adaptation logic + risk level' is output, the increment training of the increment training unit trains more than or equal to 5000 automatic line samples of the oil sprayer in each round, the termination condition is that the continuous three-wheel prediction qualification rate is more than or equal to 90%, and finally the AI parameter recommendation model with the industrial scene parameter recommendation capability is generated.
  5. 5. The intelligent parameter recommendation system suitable for the automatic line adjustment procedure of the oil sprayer according to claim 1, wherein the intelligent Agent module comprises a parameter evaluation unit for outputting a parameter adaptation degree score and a parameter fine adjustment unit for generating a step fine adjustment scheme, the parameter evaluation unit adopts a score of 0-100 points, wherein the score is more than or equal to 90, the score is qualified between 60 and 84, the score is inferior to 60, the parameter fine adjustment unit generates a differentiation scheme with the score of less than or equal to 10um being qualified or the score of 10-50um being inferior according to the score, and the adjustment frequency is less than or equal to 5 times.
  6. 6. The intelligent parameter recommendation system suitable for the automatic injector line adjustment procedure according to claim 3 is characterized in that the parameter execution module comprises a signal conversion unit and a data recording unit, wherein the signal conversion unit is used for writing an optimal propulsion parameter table generated by an intelligent Agent module into a preset database parameter table, triggering the automatic injector line control system to download optimal propulsion parameters corresponding to working conditions from the database parameter table, synchronizing the optimal propulsion parameters to a parameter configuration module of equipment, completing parameter configuration before the next production of the automatic injector line, and recording parameter writing time and parameter details before and after adjustment in real time by the data recording unit so as to ensure that the whole data link can be traced and the equipment is stable to operate.
  7. 7. The intelligent parameter recommendation system suitable for the automatic line adjustment procedure of the fuel injector according to claim 3 is characterized in that the feedback optimization module is in data interaction with the parameter execution module, the automatic line equipment data acquisition system of the fuel injector and the model training module, and the adjustment effect is quantitatively calculated by acquiring adjusted production data, when the yield is improved by more than or equal to 2% or the production efficiency is improved by more than or equal to 5%, the relevant data are preferentially fed back to the model training module for model incremental training and iterative optimization, and the adjustment data in other scenes are fed back according to forward exchange total weeks, so that the model is ensured to continuously adapt to the production requirement.
  8. 8. An intelligent parameter recommendation method suitable for an automatic line adjustment procedure of an oil sprayer is characterized by comprising the following steps: S1, system deployment, namely deploying the intelligent parameter recommendation system suitable for the automatic line adjustment procedure of the fuel injector according to any one of claims 2 to 7 on a server; S2, data acquisition, wherein a data acquisition module is in butt joint with an automatic line equipment data acquisition system of the oil sprayer to acquire real-time operation data of equipment, supplement and acquire historical production process data, and synchronize the multi-source data to a data processing module; s3, data processing, wherein the data processing module performs unified data format on the data transmitted by the data acquisition module, extracts key data comprising target flow, propulsion and qualification rate, generates a sample pair of input characteristic-optimal propulsion parameters, and simultaneously transmits the standardized data and the sample pair to a server storage unit for storage; S4, model training, namely training an AI model in an increment mode through a sample pair by a model training module until the prediction qualification rate reaches the standard; S5, the system operates, wherein a parameter evaluation unit of the intelligent Agent module invokes real-time data and outputs parameter adaptation degree scores, a parameter fine adjustment unit generates a fine adjustment scheme according to the scores and sends the fine adjustment scheme to an automatic line control system of the oil injector, and a parameter execution module directionally sends control signals of the automatic line control system of the oil injector to automatic line equipment of the oil injector; s6, feedback optimization And the feedback optimization module quantifies the adjustment effect, and feeds data back to the model training module to complete closed-loop optimization.
  9. 9. The intelligent parameter recommendation method for an automatic injector line adjustment process according to claim 8, wherein step S4 specifically comprises the following steps: S41, model selection, namely selecting an AI model adapting to an automatic line scene of the fuel injector by a model selection unit, and finally determining LightGBM serving as a core AI model and matching an AI large model for outputting adaptation logic and risk level to form a mixed reasoning architecture of a machine learning model and the large model; s42, prompting word configuration, namely loading a pre-stored industrial parameter optimization special prompting word template by a prompting word optimization unit, wherein the prompting word comprises a fuel injector adjustment procedure process principle, input parameters, current working condition parameters and output format specifications, and explicitly outputting an 'optimizing propulsion quantity, an adaptation logic and a risk grade'; S43, training the model, namely training an incremental training unit according to the standard of automatic line samples of the fuel injector with the quantity of more than or equal to 5000 times, carrying out 50 times of training, using 1 ten thousand sample pairs for each time, adopting a cross entropy loss function, setting the learning rate to be 0.001 initially, attenuating 50% every 10 times, stopping training when the model prediction qualification rate of continuous 3 times is more than or equal to 90%, and finally generating an AI parameter recommendation model with the industrial scene parameter recommendation capability, wherein the model parameter adjustment logic is basically consistent with the level of an excellent engineer.
  10. 10. The intelligent parameter recommendation method suitable for use in an automatic injector line adjustment process according to claim 9, wherein step S5 specifically comprises the following steps: S51, generating a candidate parameter table, namely inquiring a database table of an automatic line equipment data acquisition system of the oil sprayer by an intelligent Agent module, acquiring all-condition production data and an input characteristic-candidate propulsion amount data set, and calling a trained AI parameter recommendation model to generate a plurality of groups of candidate propulsion amount parameter tables of covering equipment and product types in batches; S52, full-flow simulation deduction, wherein an intelligent Agent module invokes a production flow simulation engine, respectively simulates the complete process of single flow adjustment of a plurality of groups of candidate parameter tables by combining real-time production data, dynamically calculates the flow standard reaching range, the adjustment response speed and the parameter stability key index of each group of parameter tables, and forms a simulation result data set; S53, the evaluation and fine adjustment scheme is generated, wherein the parameter evaluation unit calculates the comprehensive evaluation index of each group of candidate parameter tables according to the simulation result data set and the weight of the qualification rate of 60 percent+the stability of 20 percent, and the parameter adaptation degree score is output by adopting a scoring standard of 0-100 points; S54, determining and executing an optimal parameter table, namely, an intelligent Agent module selects a parameter table with highest comprehensive score and optimal qualification rate as a final optimal propulsion amount parameter table, a signal conversion unit of a parameter execution module writes the parameter table into a preset database parameter table according to a compatible format of an automatic line control system of the oil sprayer, and triggers the control system to download optimal propulsion amount parameters corresponding to working conditions from the database parameter table, and the optimal propulsion amount parameter table is synchronized to a parameter configuration module of equipment, so that parameter configuration is completed before the automatic line of the oil sprayer is produced next time; The specific process of the step S6 is as follows: s61, effect quantification, namely collecting mass production data after the equipment executes an optimal parameter table by a feedback optimization module, quantitatively calculating an adjustment effect, and counting key index change conditions; S62, data screening and feedback, wherein when the qualification rate is improved by more than or equal to 2% or the production efficiency is improved by more than or equal to 5%, the feedback optimization module preferentially feeds back 'all-condition production data + optimal parameter table execution effect data' to the model training module; s63, iterating the model and the parameter table, namely performing incremental training by a model training module based on feedback data, optimizing a mapping relation of working condition and propulsion, generating an iterated optimal propulsion parameter table, writing the iterated optimal propulsion parameter table into a database by a parameter executing module in a next non-production period, completing continuous optimization of the parameter table, and adapting to the working condition change of a production line.

Description

Intelligent parameter recommendation system and method suitable for automatic line adjustment procedure of oil injector Technical Field The invention relates to the field of intelligent control of industrial production, in particular to an intelligent parameter recommendation system applied to an automatic line adjustment procedure and a flow adjustment procedure of an oil sprayer. Background In the automatic production of the oil sprayer, gap adjustment and flow adjustment are key core procedures for guaranteeing the quality of products, and core parameters, especially equipment propulsion parameters, directly determine the qualification rate and the production efficiency of the oil sprayer. In the prior art, an expert experience parameter table look-up method is commonly adopted to combine with a high-precision sensor for parameter adjustment, wherein an excellent engineer establishes a fixed parameter table look-up based on long-term practical experience, and on-site staff matches corresponding propulsion parameters according to data acquired by the sensor. However, this approach suffers from the following significant drawbacks: (1) Excellent engineering teachers and resources are scarce, 24-hour stay line support cannot be realized, and the propulsion parameters are difficult to update in time when facing sudden working conditions, so that the production line is easy to deviate from an optimal running state; (2) The parameter comparison table is a fixed threshold matching mode, has poor adaptability to complex and changeable non-standard working conditions in the automatic production process of the fuel injector, and has insufficient adjustment flexibility; (3) The labor cost is high, the labor maintenance parameter comparison table is required to be continuously input, the on-site parameter adjustment requirement is responded, and the manual adjustment efficiency is limited; (4) The existing data acquisition system and the engineer debugging experience do not form a complete intelligent optimization closed loop, and the value of mass production data accumulated by an automatic line of the oil sprayer is not fully released. Disclosure of Invention The technical problem to be solved by the invention is to provide an intelligent parameter recommendation system and method suitable for an automatic line adjustment procedure of an oil sprayer, so as to improve the flexibility and accuracy of parameter adjustment, reduce the labor cost and strengthen the running stability and the product quality of a production line. The intelligent parameter recommendation system suitable for the automatic line adjustment procedure of the oil sprayer comprises a data acquisition module, a data processing module, a model training module, an intelligent Agent module and a parameter execution module: The data acquisition module is used for docking the data acquisition system of the automatic line equipment of the oil sprayer so as to acquire real-time operation data of the equipment and historical production process data, and the acquired multi-source data are structured and synchronized to the data processing module; the data processing module is used for carrying out standardized processing on the multi-source data, extracting key influence factors related to propulsion parameter optimization, generating an input characteristic-optimal propulsion parameter sample pair, and simultaneously transmitting the standardized data and the standardized samples to the server storage unit for storage; The model training module is in data interaction with the server storage unit and is used for calling the standardized data and sample pairs from the server storage unit, selecting an AI model suitable for an industrial scene, performing incremental training and iterative optimization, enabling a model deep learning engineer to tune experience, equipment operation rules and mapping relation between parameters and process effects, and finally generating an AI parameter recommendation model with industrial scene parameter recommendation capability; the intelligent Agent module is used for carrying out data interaction with the model training module and the server storage unit and serving as an interaction and decision-making execution carrier of the AI parameter recommendation model, and is used for calling the standardized historical production process data from the server storage unit, calling the trained AI parameter recommendation model for reasoning, outputting the optimal propulsion parameters and the parameter adaptation degree score, and generating a standardized optimal propulsion parameter table; and the parameter execution module is in data interaction with the intelligent Agent module and the automatic line control system of the oil injector, and is used for writing an optimal propulsion parameter table into a database parameter table, triggering the automatic line control system of the oil injector to download the optimal propuls