A method and system for controlling energy consumption in aluminum profile extrusion process

Abstract

The present invention relates to a method and system for controlling energy consumption during an aluminum extrusion process. The system constructs a control network with an industrial computer as the host computer and controllers for the extruder, aluminum bar heating furnace, and mold heating furnace equipment as slave computers. Main control software is designed for the industrial computer, and communication between the host computer and slave equipment is established via the TCP / IP protocol. The system also exchanges data with a manufacturing execution environment (MES) system and acquires real-time information. Process parameters during the aluminum extrusion process are obtained from the MES system, and the real-time data of these process parameters is input into a machine learning model to output optimized process parameters. Based on these process parameters, the software module automatically and accurately calculates the preheating temperatures of the mold and aluminum bar, as well as the opening times of the mold heating furnace door and the aluminum bar heating furnace door, for energy consumption control. By combining the industrial computer, MES system, and machine learning technology, as well as accurately calculating equipment control parameters, efficient energy consumption control and production optimization are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy consumption control in aluminum profile extrusion production, and in particular to an energy consumption control method and system for an aluminum profile extrusion process. Background Art

[0002] The rapid development of the global economy has led to the rapid development of fields such as construction, industry, aerospace, etc. Due to the excellent material ductility, easy processing, and formability of aluminum profiles, the market demand for them in these fields remains high.

[0003] The aluminum extrusion process not only requires an aluminum rod heating furnace and a mold heating furnace to preheat the aluminum rod and mold, respectively, but also requires the extruder to use a hydraulic system to generate a very high extrusion force on the extrusion rod to extrude the aluminum rod through the corresponding mold to obtain the desired shape of the product. Therefore, aluminum profile extrusion production is a high-energy, high-emission production process. However, the aluminum rod heating furnace, mold heating furnace, extruder, and MES system in the extrusion workshop of aluminum profile manufacturers are currently relatively independent and not integrated to provide unified control of the aluminum profile extrusion process. As a result, various process step information cannot be used to control each device, and the order data from the MES system cannot be directly used to generate control parameters for extrusion production. Parameters such as the aluminum rod preheating temperature, mold preheating temperature, and bar length are mostly calculated manually, and the opening time of the aluminum rod heating furnace door and the mold heating furnace door is controlled based on experience.

[0004] Therefore, in order to ensure that the aluminum bar temperature and mold temperature meet the requirements of the extrusion process, the aluminum bar preheating temperature and mold preheating temperature are often increased. This not only leads to the aluminum bar preheating temperature and mold preheating temperature being too high, and being stranded outside the equipment, that is, the process from leaving the heating furnace to installing it on the extruder is too long, which wastes too much energy, fails to achieve energy conservation and emission reduction in aluminum profile extrusion production, but also increases the labor intensity of workers.

[0005] Therefore, in the context of carbon peak and carbon neutrality, how to improve production efficiency, control production energy consumption and minimize energy consumption through system integration and precise calculation of equipment control parameters is still a technical problem that needs to be solved urgently in aluminum profile extrusion production. Summary of the Invention

[0006] The purpose of the present invention is to provide an energy consumption control method and system for an aluminum profile extrusion process, which improves production efficiency by integrating the system and accurately calculating the control parameters of each device to achieve minimized energy consumption.

[0007] According to one main aspect of the present invention, a method for controlling energy consumption in an aluminum profile extrusion process is provided, comprising one or a combination of the following steps:

[0008] Step 1: Build a control network with an industrial computer as the host computer and controllers of the extruder, aluminum bar heating furnace, and mold heating furnace equipment as slave computers; the control network adopts the industrial Ethernet communication protocol and realizes real-time data transmission and the issuance of control instructions;

[0009] Step 2: Design the main control software for the industrial computer, establish communication between the host computer and the slave devices through the TCP / IP protocol, and exchange data and obtain real-time information with the MES system;

[0010] Step 3: Obtain the process parameters of the aluminum profile extrusion process from the MES system, including the mold temperature T m , aluminum bar extrusion temperature T j , aluminum rod length L, extrusion speed v j , Extrusion rod idle stroke S k , Extrusion rod no-load forward speed v kq , Extrusion rod no-load retreat v kh ; Input the real-time data of the process parameters into the machine learning model and output the optimized process parameters;

[0011] Step 4: Test the temperature drop rate △T of the preheated aluminum bar and preheated mold at room temperature lb , △T m ;

[0012] Step 5: Test the time it takes for the mold to be taken out of the mold heating furnace and installed on the extruder. mjaz ;

[0013] Step 6: Test the aluminum rod cutting time t cj , the time it takes for the aluminum bar to be installed on the extruder t lbaz ;

[0014] Step 7: Calculate the working time t of one aluminum bar extrusion cycle jxh The working time of one aluminum bar extrusion cycle is t jxh is the extrusion time of a single aluminum bar t jy , Extrusion rod no-load retreat time t kh , Extrusion rod no-load forward time t kq sum;

[0015] Among them, the extrusion time of the single aluminum rod is The no-load retreat time of the extrusion rod

[0016] The time for the extrusion rod to move forward without load

[0017] Step 8: Calculate the time t that the aluminum bar remains outside the equipment lbzl The length of time t that the aluminum rod is retained outside the devicelbzl The time t for the aluminum rod to be pushed out of the heating furnace tc , aluminum bar cutting time t cj , the time it takes for the aluminum bar to be installed on the extruder t lbaz sum;

[0018] The time it takes for the aluminum rod to be pushed out of the heating furnace is Where L0 is the distance from the aluminum bar heating furnace door to the shear knife, v lbtc The speed at which the aluminum rod is pushed out of the heating furnace.

[0019] Step 9: Calculate the time t in the extrusion cycle corresponding to the time when the aluminum bar heating furnace door is opened based on the obtained parameters. lblm The time t in the extrusion cycle corresponding to the time when the mold heating furnace door is opened is mjlm ;

[0020] Step 10: Calculate the required preheating temperature T of the aluminum bar based on the preset production requirements and extrusion process parameters. lby And the mold preheating temperature T my ;

[0021] Step 11: Control the aluminum rod heating furnace and the mold heating furnace according to the calculated parameters to preheat the aluminum rod and the mold to the temperature calculated in step 10, and open the aluminum rod heating furnace door and the mold heating furnace door at the time point calculated in step 9.

[0022] In the above solution, efficient energy consumption control and production optimization are achieved through the combination of industrial computers, MES systems and machine learning technology, system integration and precise calculation of equipment control parameters. First, a control network is established between the host computer (industrial computer) and the slave computers (extrusion press, aluminum bar heating furnace, and mold heating furnace controller) using the Industrial Ethernet communication protocol. Compared with traditional Ethernet, Industrial Ethernet is more suitable for industrial environments and has high reliability, real-time performance and anti-interference capabilities. Through Industrial Ethernet, seamless communication between devices can be achieved, ensuring the coordinated operation of equipment such as extruders and heating furnaces, reducing production inconsistencies caused by information delays, improving the response speed and accuracy of the production line, and ensuring that each device can respond to operating instructions in the shortest possible time.

[0023] Designing master control software on an industrial computer uses TCP / IP to communicate with lower-level devices and simultaneously exchange data and acquire real-time information with the MES system. Through integration with the MES system, the master control software provides comprehensive monitoring of the production process, ensuring real-time synchronization of production parameters with order requirements.

[0024] Furthermore, through the introduction of machine learning models, process parameter optimization no longer relies on empirical experience, but rather on dynamic analysis of historical data and real-time operating conditions. The model enables continuous learning and optimization during the production process, enhancing the level of intelligent production. By analyzing energy consumption data in real time, the model predicts and optimizes production conditions to minimize energy consumption. Energy waste is reduced by automatically adjusting parameters such as heating temperature and extrusion speed.

[0025] By testing the installation and preheating process of aluminum bars and molds, the time for each operation link is accurately calculated, and the opening time of the heating furnace door is adjusted according to real-time data, so that energy consumption in the production process is minimized.

[0026] In some embodiments, as a further preferred embodiment, the mold temperature T m , aluminum bar extrusion temperature T j Real-time monitoring using thermocouple sensors. Thermocouple sensors can maintain stable performance in high-temperature environments and provide rapid temperature feedback, helping to adjust the operation of the heating furnace and extruder in a timely manner to ensure that the process temperature is always within the optimal range.

[0027] In some embodiments, as a further preferred embodiment, the aluminum rod heating furnace door is opened at the time t in the extrusion cycle. lblm The calculation formula is:

[0028] When mold replacement is not required: t lblm =t jxh -t lbzl ,

[0029] When mold replacement is required: t lblm =t jxh +t mjaz -t lbzl ,

[0030] The mold heating furnace door is opened after the last product of a batch is completed. The opening time is t mjlm =t jxh -t mjaz Specifically, this heating furnace door opening control based on the extrusion cycle and production rhythm greatly improves the continuity and efficiency of the production process and reduces unnecessary energy waste.

[0031] In some embodiments, as a further preferred embodiment, the preheating temperature required for the aluminum rod is The preheating temperature T required for the mold my =T m +△T m ×t mjaz .

[0032] In some embodiments, as a further preferred embodiment, the mold temperature T m , aluminum bar extrusion temperature T j Real-time monitoring is performed using thermocouple sensors.

[0033] In some embodiments, as a further preferred solution, the main control software of the industrial computer is provided with a timer for the opening time of the aluminum bar heating furnace door;

[0034] The aluminum rod heating furnace door opening time timer starts timing from 0 when the aluminum rod begins to be extruded, that is, when the pressure sensor of the hydraulic cylinder of the extruder detects that the cylinder pressure is greater than the hydraulic cylinder pressure when the extrusion rod moves forward without load.

[0035] In some embodiments, as a further preferred embodiment, when the aluminum rod heating furnace door opening timer counts to the time when the aluminum rod heating furnace door opens, the main control software sends an instruction to the controller of the aluminum rod heating furnace, the aluminum rod heating furnace door opens, and the aluminum rod is pushed out of the heating furnace, and then cut according to the length of the aluminum rod and installed on the extruder.

[0036] The technological innovation of incorporating a timer to indicate when the aluminum bar heating furnace door opens within the industrial computer's main control software effectively optimizes heating control and energy management during aluminum extrusion production. By triggering the timer based on real-time data from the extruder's hydraulic cylinder pressure sensor and precisely controlling the opening of the heating furnace door, the system ensures that the aluminum bars enter the extruder at the optimal temperature, minimizing the impact of temperature fluctuations on product quality. This intelligent control not only improves the stability and consistency of the production process but also significantly reduces energy consumption, providing crucial technical support for enterprises to achieve efficient and energy-efficient production.

[0037] Moreover, after the aluminum rod heating furnace door opening time timer counts to the time when the aluminum rod heating furnace door is opened, the main control software sends an opening instruction to the aluminum rod heating furnace controller, thereby realizing the automated process of smoothly pushing the aluminum rod out of the heating furnace and accurately cutting and installing it into the extruder according to the length of the aluminum rod.

[0038] In some embodiments, as a further preferred solution, a product counter and a mold heating furnace door opening time timer are set in the main control software of the industrial computer; when the count of the product counter is equal to the number of product batches - 1 and the aluminum rod starts to be extruded again, the mold heating furnace door opening time timer starts counting from 0.

[0039] In some embodiments, as a further preferred embodiment, when the mold heating furnace door opening timer counts to the time when the mold heating furnace door opens, the main control software sends an instruction to the controller of the mold heating furnace, the mold heating furnace door opens, and the mold is taken out of the heating furnace and installed on the extruder.

[0040] In some embodiments, a product counter accurately records the number of products in each batch and triggers a timer to indicate when the mold door opens, starting from zero, when aluminum bar extrusion begins. By precisely recording the moment the mold door opens, the main control software can issue instructions at the optimal time, ensuring that the mold is removed and installed in the extruder immediately after reaching the desired temperature. This precise control of the mold door opening timing avoids unnecessary energy loss within the furnace.

[0041] In some embodiments, as a further preferred solution, the machine learning model is trained in combination with historical process parameters and production data of the MES system; the trained machine learning model is exported and stored in an industrial computer; a machine learning module is set in the main control software, and the trained model is loaded to ensure that the module can access real-time process parameter data;

[0042] The machine learning model includes but is not limited to supervised learning algorithms, unsupervised learning algorithms, and reinforcement learning algorithms.

[0043] In the above solution, the machine learning model is trained using historical data from the MES system to learn the impact of different process parameters on energy consumption and production efficiency. Before entering the data into the machine learning model, data preprocessing is required to ensure data quality and availability, and to normalize or standardize data features so that different features do not cause deviations due to different dimensions during model training. This data is divided into training, validation, and test sets, with 80% of the data used for model training, 10% for validation, and 10% for testing. The training set is used for model learning, the validation set is used to adjust hyperparameters (such as model depth and learning rate), and the test set is used to evaluate the model's performance on unseen data.

[0044] In some embodiments, supervised learning algorithms, such as random forests or linear regression, are used to train data within an industrial computer. Based on historical data, the algorithms learn the relationship between different process parameters and energy consumption, predict their impact on energy consumption, and identify the optimal combination of production parameters. Random forests are suitable for finding the optimal combination of multiple process parameters to minimize energy consumption. Linear regression models are used to predict the relationship between specific extrusion process parameters and energy consumption.

[0045] In some embodiments, unsupervised learning algorithms are used to detect equipment anomalies or optimize performance. Clustering algorithms are used to categorize equipment operating conditions into distinct groups, helping to identify when equipment is experiencing abnormal conditions. Reinforcement learning algorithms are used to adjust and optimize production parameters in real time to minimize energy consumption and maximize production efficiency. Through continuous interaction with the environment, the reinforcement learning model can learn the optimal extrusion parameter settings to minimize energy consumption and maximize production efficiency.

[0046] In some embodiments, a reinforcement learning algorithm is used to learn optimal production parameter settings in interaction with the environment to minimize energy consumption and maximize production efficiency.

[0047] In some embodiments, the model is validated using k-fold cross-validation (typically 5-fold or 10-fold) to ensure robustness and generalization. Grid search or random search is used to adjust the model's hyperparameters to ensure the model can make accurate predictions based on MES system data.

[0048] In some embodiments, the trained machine learning model is exported to a format readable by an industrial computer and stored in the industrial computer. A machine learning module is implemented within the main control software of the industrial computer to ensure that the trained model can be loaded and executed. During the production process, the machine learning model acquires real-time process data from the MES system and generates predictions. Based on the predictions from the real-time data, the model automatically generates optimal operating parameters and transmits these parameters to the MES system.

[0049] According to another main aspect of the present invention, an energy consumption control system for an aluminum profile extrusion process is provided, which is used to implement the energy consumption control method for an aluminum profile extrusion process, and is characterized by comprising:

[0050] Software module for real-time reading and optimization of various process parameters of the MES system and calculation of various control parameters;

[0051] The equipment module mainly includes a hydraulic extruder, a die heating furnace, and an aluminum bar heating furnace, which are used to preheat the die and aluminum bar and extrude the aluminum bar;

[0052] The communication and control module is used for communication and control of equipment modules, realizing the integration and coordination of various parts of the system. It uses Ethernet communication protocol, TCP / IP protocol and control technology to achieve precise monitoring and adjustment of the extrusion process.

[0053] Advantages and beneficial effects of the present invention:

[0054] The equipment and MES system in the aluminum extrusion production workshop are no longer independent entities; instead, they are integrated into one, enabling unified automatic control of equipment and automated data transmission. Furthermore, the main control software, combined with machine learning models, provides real-time decision support and optimization suggestions based on process parameter data provided by the MES system. Based on historical data and real-time monitoring results, the system automatically adjusts production plans and parameters, making decision-making more scientific and efficient.

[0055] Secondly, various process parameters are automatically read from the MES system. By accurately calculating key parameters such as the aluminum bar preheat temperature, the die preheat temperature, the extrusion cycle duration, and the aluminum bar residence time, each step in the extrusion process is ensured to be carried out under optimal conditions. These calculations are based on real-time data and take into account the equipment's operating status and production rhythm, making process parameters such as temperature, time, and speed more precise and controllable.

[0056] Finally, automated process calculations and control systems reduce the need for human intervention. The system automatically operates based on calculations, reducing process deviations and errors caused by human factors, making the production process more stable and reliable. This also strengthens the company's competitive advantage and leading position in the market. This technological innovation provides the company with broader development space and the driving force for sustained growth. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying any creative work, other drawings obtained based on these drawings still fall within the scope of the present invention.

[0058] Figure 1 It is a flow chart of the energy consumption control method of the aluminum profile extrusion process of the present invention. DETAILED DESCRIPTION

[0059] The preferred embodiments of the present invention will be described in detail below so that the purpose, features and advantages of the present invention can be more clearly understood. It should be understood that the following embodiments are not intended to limit the scope of the present invention, but are only intended to illustrate the essential spirit of the technical solution of the present invention.

[0060] In the following description, for the purpose of illustrating the various disclosed embodiments, certain specific details are set forth in order to provide a thorough understanding of the various disclosed embodiments. However, those skilled in the relevant art will recognize that the embodiments may be practiced without one or more of these specific details. In other cases, well-known devices, structures, and techniques associated with this application may not be shown or described in detail to avoid unnecessarily obscuring the description of the embodiments.

[0061] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any manner in one or more embodiments.

[0062] The following are specific embodiments of the present invention and the accompanying drawings to further describe the technical solutions of the present invention, but the present invention is not limited to these embodiments.

[0063] like Figure 1 FIG. 1 is a flow chart of a method for controlling energy consumption during an aluminum profile extrusion process according to an embodiment of the present invention.

[0064] An embodiment of the present invention provides a method for controlling energy consumption during aluminum profile extrusion, comprising one or a combination of the following steps:

[0065] Step 1: Build a control network with an industrial computer as the host computer and controllers of the extruder, aluminum bar heating furnace, and mold heating furnace equipment as slave computers; the control network adopts the industrial Ethernet communication protocol and realizes real-time data transmission and the issuance of control instructions;

[0066] Step 2: Design the main control software for the industrial computer, establish communication between the host computer and the slave devices through the TCP / IP protocol, and exchange data and obtain real-time information with the MES system;

[0067] Step 3: Obtain the process parameters of the aluminum profile extrusion process from the MES system, including the mold temperature T m =420℃, aluminum bar extrusion temperature T j =420℃, aluminum bar length L=810mm, extrusion speed v j =3mm / s, extrusion rod idle stroke S k =300mm, extrusion rod no-load forward speed v kq =30mm / s, extrusion rod no-load retreat v kh =50mm / s; inputting the real-time data of the process parameters into the machine learning model and outputting the optimized process parameters;

[0068] Step 4: Test the temperature drop rate △T of the preheated aluminum bar and preheated mold at room temperature lb =0.5℃ / s, △T m =0.5℃ / s;

[0069] Step 5: Test the time it takes for the mold to be taken out of the mold heating furnace and installed on the extruder. mjaz =15s;

[0070] Step 6: Test the aluminum rod cutting time t cj = 6s, the time it takes for the aluminum bar to be installed on the extruder t lbaz =12s;

[0071] Step 7: Calculate the working time t of one aluminum bar extrusion cycle jxhThe working time of one aluminum bar extrusion cycle is t jxh is the extrusion time of a single aluminum rod t jy , Extrusion rod no-load retreat time t kh , Extrusion rod no-load forward time t kq sum;

[0072] Among them, the extrusion time of the single aluminum rod is The no-load retreat time of the extrusion rod The time for the extrusion rod to move forward without load

[0073] The calculation of the working time of one aluminum bar extrusion cycle t jxh =t jy +t kq +t kh =270+22.2+10=302.2s;

[0074] Step 8: Calculate the time t that the aluminum bar remains outside the equipment lbzl The length of time t that the aluminum rod is retained outside the device lbzl The time t for the aluminum rod to be pushed out of the heating furnace tc , aluminum bar cutting time t cj , the time it takes for the aluminum bar to be installed on the extruder t lbaz sum;

[0075] The distance from the aluminum bar heating furnace door to the shear knife is L0 = 500 mm. The speed at which the aluminum bar is pushed out of the heating furnace is v. lbtc =40mm / s, where the time it takes for the aluminum rod to be pushed out of the heating furnace

[0076] The length of time t during which the aluminum rod remains outside the equipment lbzl t lbzl =t tc +t cj +t lbaz =32.75+6+12=50.75s;

[0077] Step 9: Calculate the time t in the extrusion cycle corresponding to the time when the aluminum bar heating furnace door is opened based on the obtained parameters. lblm The time t in the extrusion cycle corresponding to the time when the mold heating furnace door is opened is mjlm ;

[0078] When mold replacement is not required, the aluminum bar heating furnace door is opened at t lblm =t jxh -t lbzl =302.2-50.75=251.45s;

[0079] When the mold needs to be replaced, the aluminum bar heating furnace door is opened at t lblm =t jxh +t mjaz -t lbzl =302.2+15-50.75=266.45s;

[0080] The time when the mold heating furnace door is opened corresponds to the time t in the extrusion cycle mjlm =t jxh -t mjaz =302.2-15=287.2s;

[0081] Step 10: Calculate the required preheating temperature T of the aluminum bar based on the preset production requirements and extrusion process parameters. lby And the mold preheating temperature T my ;

[0082]

[0083] T my =T m +△T m ×t mjaz =420+0.5×15=437.5℃;

[0084] Step 11: Control the aluminum rod heating furnace and the mold heating furnace according to the calculated parameters to preheat the aluminum rod and the mold to the temperature calculated in step 10, and open the aluminum rod heating furnace door and the mold heating furnace door at the time point calculated in step 9.

[0085] As a further preferred solution, the mold temperature T m , aluminum bar extrusion temperature T j Real-time monitoring is performed using thermocouple sensors.

[0086] As a further preferred embodiment, a timer for the aluminum bar heating furnace door opening is implemented within the main control software of the industrial computer. This timer begins counting from 0 when the aluminum bar begins to be extruded, i.e., when the pressure sensor of the extruder's hydraulic cylinder detects that the cylinder pressure is greater than the pressure of the extrusion rod when it is advancing unloaded. When the time reaches the aluminum bar heating furnace door opening time, the main control software issues a command to the aluminum bar heating furnace controller, causing the aluminum bar heating furnace door to open and the aluminum bar to be ejected from the furnace. The aluminum bar is then cut to length and mounted on the extruder.

[0087] As a further preferred embodiment, a product counter and a mold heating furnace door opening timer are configured in the main control software of the industrial computer. When the product counter reaches the number of product batches minus 1 and aluminum bar extrusion resumes, the mold heating furnace door opening timer starts counting from 0. When the mold heating furnace door opening timer is reached, the main control software issues a command to the mold heating furnace controller, causing the mold heating furnace door to open, and the mold is removed from the furnace and installed on the extruder.

[0088] In some embodiments, before the process parameter data is input into the machine learning model, data preprocessing is required to ensure the quality and availability of the data and to normalize or standardize the data features so that different features will not produce deviations due to different dimensions during model training.

[0089] In some embodiments, the machine learning algorithm is trained in combination with historical process parameters and production data of the MES system; the trained machine learning model is exported and stored in an industrial computer; a machine learning module is set up in the main control software, and the trained model is loaded to ensure that the module can access the real-time data of the process parameters;

[0090] The machine learning algorithm includes but is not limited to a supervised learning algorithm, an unsupervised learning algorithm, and a reinforcement learning algorithm.

[0091] The machine learning model is trained using historical data from the MES system to learn how different process parameters affect energy consumption and production efficiency. This data is divided into training, validation, and test sets, with 80% of the data used for model training, 10% for validation, and 10% for testing. The training set is used to learn the model, the validation set is used to adjust hyperparameters (such as model depth and learning rate), and the test set is used to evaluate the model's performance on unseen data.

[0092] As a further preferred option, supervised learning algorithms such as random forests or linear regression are used to train data within industrial computers. Based on historical data, the algorithms learn the relationship between different process parameters and energy consumption, predict their impact on energy consumption, and identify the optimal combination of production parameters. Random forests are suitable for finding the optimal combination of multiple process parameters to minimize energy consumption. Linear regression models are used to predict the relationship between specific extrusion process parameters and energy consumption.

[0093] Random forest is an ensemble learning algorithm that improves the accuracy and stability of the model by building multiple decision trees.

[0094] In some embodiments, the training process of the random forest ensemble learning algorithm:

[0095] Using the bootstrap method, different subsets are extracted from the training set, and each subset is used to train a decision tree. Each tree partitions the data according to specific process parameter selection criteria, making the samples at each leaf node as similar as possible (partitioning is done by minimizing the Gini coefficient or information gain). A voting mechanism is used to determine the final result based on the predictions of each tree. In regression problems, the prediction result is the average of all trees.

[0096] The random forest model learns the impact of process parameters on energy consumption through a voting mechanism of multiple trees and predicts the optimal combination of process parameters to minimize energy consumption.

[0097] For regression problems, the mean square error is used as the loss function to measure the deviation between the predicted value and the true value. The MSE formula is as follows:

[0098] Among them, y i is the actual value, y^ i is the model prediction value, and n is the number of samples.

[0099] The random forest model is able to calculate the importance of each feature and determine which process parameters have the greatest impact on energy consumption by looking at the average information gain or the decrease in the Gini coefficient when splitting.

[0100] Linear regression is used to analyze the linear relationship between specific process parameters and energy consumption.

[0101] In some embodiments, the training process of the linear regression learning algorithm:

[0102] Linear regression assumes a linear relationship between energy consumption Y and process parameters X. The model form is: Y = + β0 + β1X1 + β2X2 + ... + β p X p +∈

[0103] Among them, β0 is the intercept, β2,...,β p , is the weight of the process parameters, ∈ is the error term.

[0104] The goal of the model is to find the optimal β parameter by minimizing the error, using the least squares method: That is, minimize the sum of squares of actual and predicted values.

[0105] After obtaining the optimal parameters, new process parameters can be input to predict their impact on energy consumption.

[0106] As a further preferred solution, unsupervised learning algorithms are used to detect equipment anomalies or optimization. Clustering algorithms are used to group equipment operating states into different groups, helping to identify when equipment is experiencing abnormal conditions. Reinforcement learning algorithms are used to adjust and optimize production parameters in real time to achieve the lowest energy consumption and the highest production efficiency. Through continuous interaction with the environment, the reinforcement learning model can learn the optimal extrusion parameter settings to minimize energy consumption and maximize production efficiency.

[0107] K-means clustering is a typical unsupervised learning algorithm used to classify equipment operating status data to identify whether the equipment is in an abnormal state.

[0108] Select K initial centroids (representing different running state clusters).

[0109] For each sample, calculate its Euclidean distance to each centroid and assign the sample to the closest cluster: Among them, d(x i ,c k ) is the sample x i to cluster centroid c k distance.

[0110] Update centroids: The new centroid of each cluster is the average of all samples in the cluster.

[0111] Repeat the above process until the centroid no longer changes significantly (i.e., the clustering results converge).

[0112] Through clustering algorithms, the operating status of the equipment can be divided into normal and abnormal groups, thereby identifying whether the equipment is abnormal.

[0113] As a further preferred solution, the optimal production parameter settings are learned through interaction with the environment through reinforcement learning algorithms to minimize energy consumption and maximize production efficiency.

[0114] State S: current production process parameters; Action A: the action of adjusting production parameters; Reward R: feedback given by the system after the action is executed; Strategy π: decision-making rules or strategies that determine what actions to take in a certain state.

[0115] Calculation process:

[0116] Q-table initialization: Initialize a Q-value table for each state-action pair.

[0117] Policy update: Based on the current state, an action is chosen (which can be a greedy choice based on maximizing the current Q-value or a random choice that explores other actions).

[0118] Execute actions and get feedback: After the action is executed, the environment gives a reward (positive or negative) and enters the next state.

[0119] Q value update:

[0120] Among them, α is the learning rate, which is a parameter between 0 and 1. In the early stage of the model, an α value of 0.5 can accelerate learning. During the training process, a learning rate decay strategy is used, that is, the learning rate is gradually reduced from 0.1 to 0.01 as the training progresses;

[0121] γ is a discount factor that determines the importance of future rewards in current decisions and measures the degree of discount of future rewards. Its value is 0.9 or 0.95.

[0122] R is the immediate reward, which is the feedback value returned by the environment to the agent after the agent takes a specific action in a certain state. It indicates how "good" or "bad" the action is in the current state. Positive rewards are given to successful actions to encourage the agent to take such actions. Negative rewards are given to failures or bad actions to inhibit such actions.

[0123] Q(s′,a′) is the optimal Q value of the next state.

[0124] Through repeated interactions, reinforcement learning gradually optimizes the Q table and finds the combination of production parameters that minimizes energy consumption and maximizes production efficiency.

[0125] As a further preferred approach, the model is validated using K-fold cross-validation (typically 5-fold or 10-fold) to ensure robustness and generalization. K-fold cross-validation divides the data into K subsets, cyclically using one of these subsets as the validation set and the other K-1 subsets as the training set. The validation error of the model is calculated K times, and the average is taken as the final error evaluation metric.

[0126] As a further preferred approach, grid search or random search is used to adjust the model's hyperparameters to ensure that the model can make accurate predictions based on MES system data. Grid search searches for the optimal combination of hyperparameters within a predefined grid. Random search randomly samples the hyperparameter space several times to find the most effective hyperparameter combination.

[0127] Export the trained machine learning model to a format readable by an industrial computer and store it there. Set up a machine learning module in the main control software of the industrial computer to ensure that the trained model can be loaded and run. During production, the machine learning model obtains real-time process data from the MES system and generates predictions. Based on the predictions from this real-time data, the model automatically generates optimal operating parameters and transmits these parameters to the MES system.

[0128] As a further preferred solution, at each stage of the extrusion process, current process parameters are collected in real time and input into the model to predict the optimal heating temperature and extrusion time for the aluminum bar and die. Based on the model's predictions, the temperature setpoints of the aluminum bar and die heating furnaces are dynamically adjusted. At the beginning of the extrusion process, the extrusion speed is adjusted based on real-time temperature and pressure information to ensure product quality.

[0129] When the model predicts that the set aluminum bar heating temperature is about to be reached, the main control software sends a command to the lower computer device through the TCP / IP protocol to open the aluminum bar heating furnace door.

[0130] The model's real-time prediction capabilities were tested in an actual production environment, and the deviation between the model's output and actual production efficiency was monitored. Model parameters and control strategies were continuously adjusted based on real-time monitoring data to optimize the heating time and temperature of the aluminum bars and molds.

[0131] By integrating machine learning algorithms into industrial computer control software, smarter control of the aluminum extrusion process can be achieved.

[0132] Based on the second main aspect of the present invention, an energy consumption control system for an aluminum profile extrusion process is provided, which is used to implement the energy consumption control method for an aluminum profile extrusion process, comprising:

[0133] Software module for real-time reading and optimization of various process parameters of the MES system and calculation of various control parameters;

[0134] The equipment module mainly includes a hydraulic extruder, a die heating furnace, and an aluminum bar heating furnace, which are used to preheat the die and aluminum bar and extrude the aluminum bar;

[0135] The communication and control module is used for communication and control of equipment modules, realizing the integration and coordination of various parts of the system. It uses Ethernet communication protocol, TCP / IP protocol and control technology to achieve precise monitoring and adjustment of the extrusion process.

[0136] Anything not described in detail in the present invention is well known to those skilled in the art.

[0137] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for controlling energy consumption in an aluminum profile extrusion process, characterized in that: The method includes one or a combination of the following steps: Step 1: Build a control network with an industrial computer as the host computer and controllers of the extruder, aluminum bar heating furnace, and mold heating furnace as slave computers; the control network uses the industrial Ethernet communication protocol to achieve real-time data transmission and the issuance of control instructions; Step 2: Design the main control software for the industrial computer, establish communication between the host computer and the slave devices through the TCP / IP protocol, and exchange data and obtain real-time information with the MES system; Step 3: Obtain the process parameters of the aluminum profile extrusion process from the MES system, including the mold temperature T m , aluminum bar extrusion temperature T j , aluminum rod length L, extrusion speed v j , Extrusion rod idle stroke S k , Extrusion rod no-load forward speed v kq 、Extrusion rod no-load retreat v kh ; Input the real-time data of the process parameters into the machine learning model and output the optimized process parameters; Step 4: Test the temperature drop rate △T of the preheated aluminum bar and preheated mold at room temperature lb , △T m ; Step 5: Test the time it takes for the mold to be taken out of the mold heating furnace and installed on the extruder. mjaz ; Step 6: Test the aluminum rod cutting time t cj , the time it takes for the aluminum bar to be installed on the extruder t lbaz ; Step 7: Calculate the working time t of one aluminum bar extrusion cycle jxh The working time of one aluminum bar extrusion cycle is t jxh is the extrusion time of a single aluminum rod t jy , Extrusion rod no-load retreat time t kh , Extrusion rod no-load forward time t kq sum; Among them, the extrusion time of the single aluminum rod is The no-load retreat time of the extrusion rod The time for the extrusion rod to move forward without load Step 8: Calculate the time t that the aluminum bar remains outside the equipment lbzl The aluminum rod is retained outside the device for a period of time t lbzl The time t for the aluminum rod to be pushed out of the heating furnace tc , aluminum bar cutting time t cj , the time it takes for the aluminum bar to be installed on the extruder t lbaz sum; The time it takes for the aluminum rod to be pushed out of the heating furnace is Where L0 is the distance from the aluminum bar heating furnace door to the shear knife, v lbtc The speed at which the aluminum rod is pushed out of the heating furnace; Step 9: Based on the obtained parameters, calculate the time t in the extrusion cycle when the aluminum bar heating furnace door is opened. lblm The time t in the extrusion cycle corresponding to the time when the mold heating furnace door is opened is mjlm ; Step 10: Calculate the required preheating temperature T of the aluminum bar based on the preset production requirements and extrusion process parameters. lby And the mold preheating temperature T my ; Step 11: Control the aluminum rod heating furnace and the mold heating furnace according to the calculated parameters to preheat the aluminum rod and the mold to the temperature calculated in step 10, and open the aluminum rod heating furnace door and the mold heating furnace door at the time point calculated in step 9.

2. The method for controlling energy consumption during aluminum extrusion according to claim 1, wherein: The mold temperature T m , aluminum bar extrusion temperature T j Real-time monitoring is performed using thermocouple sensors.

3. The method for controlling energy consumption during aluminum extrusion according to claim 1, wherein: The aluminum rod heating furnace door is opened at the time t in the extrusion cycle. lblm The calculation formula is: When mold replacement is not required: t lblm =t jxh -t lbzl , When mold replacement is required: t lblm =t jxh +t mjaz -t lbzl , The mold heating furnace door is opened after the last product of a batch is completed. The opening time is t mjlm =t jxh -t mjaz .

4. The method for controlling energy consumption during aluminum profile extrusion according to claim 1, wherein: The preheating temperature required for the aluminum rod The preheating temperature T required for the mold my =T m +△T m ×t mjaz .

5. The method for controlling energy consumption during aluminum profile extrusion according to claim 1, wherein: The main control software of the industrial computer is provided with a timer for the opening time of the aluminum bar heating furnace door; The aluminum rod heating furnace door opening time timer starts timing from 0 when the aluminum rod begins to be extruded, that is, when the pressure sensor of the hydraulic cylinder of the extruder detects that the cylinder pressure is greater than the hydraulic cylinder pressure when the extrusion rod moves forward without load.

6. The method for controlling energy consumption during aluminum profile extrusion according to claim 5, characterized in that: When the aluminum rod heating furnace door opening time timer counts to the time when the aluminum rod heating furnace door is opened, the main control software sends an instruction to the controller of the aluminum rod heating furnace, the aluminum rod heating furnace door opens, the aluminum rod is pushed out of the heating furnace, and then cut according to the length of the aluminum rod and installed on the extruder.

7. The method for controlling energy consumption during aluminum extrusion process according to claim 1, characterized in that: The main control software of the industrial computer is provided with a product counter and a mold heating furnace door opening time timer; when the count of the product counter is equal to the number of product batches minus 1 and the aluminum bar starts to be extruded again, the mold heating furnace door opening time timer starts counting from 0.

8. The method for controlling energy consumption during aluminum profile extrusion according to claim 7, characterized in that: When the mold heating furnace door opening time timer counts to the time when the mold heating furnace door is opened, the main control software sends an instruction to the controller of the mold heating furnace, the mold heating furnace door opens, and the mold is taken out of the heating furnace and installed on the extruder.

9. The method for controlling energy consumption during aluminum profile extrusion according to claim 1, wherein: The machine learning model is trained in combination with historical process parameters and production data from the MES system; the trained machine learning model is exported and stored in an industrial computer; a machine learning module is set up in the main control software, and the trained model is loaded to ensure that the module can access real-time process parameter data; The machine learning model includes but is not limited to supervised learning algorithms, unsupervised learning algorithms, and reinforcement learning algorithms.

10. An energy consumption control system for an aluminum profile extrusion process, used to implement the energy consumption control method for an aluminum profile extrusion process according to any one of claims 1 to 8, characterized in that: include: Software module for real-time reading and optimization of various process parameters of the MES system and calculation of various control parameters; The equipment module mainly includes a hydraulic extruder, a die heating furnace, and an aluminum bar heating furnace, which are used to preheat the die and aluminum bar and extrude the aluminum bar; The communication and control module is used for communication and control of equipment modules, realizing the integration and coordination of various parts of the system. Through industrial Ethernet communication protocol, TCP / IP protocol and control technology, it can realize accurate monitoring and adjustment of the extrusion process.

Images