A high-efficiency energy-consuming extrusion speed adjustment system and method

Abstract

The present invention discloses an efficient and energy-saving extrusion speed adjustment system and method, which relates to the field of automated control technology. When the system is in operation, the accuracy and timeliness of the data are ensured by collecting the operating status information of the extruder equipment in real time. Through in-depth feature extraction and screening of the collected data, the system can build a high-precision prediction model, and predict potential abnormal situations in advance through machine learning algorithms. When possible risks are predicted, dynamic load optimization is automatically triggered, and the extrusion speed and pressure are intelligently adjusted to achieve the optimization of production efficiency and energy efficiency. In addition, through real-time PLC and servo motor control, the stability and efficiency of the extrusion process are ensured, which significantly reduces the dependence on manual intervention and reduces the risk of operational errors. It overcomes the shortcomings of response lag and dependence on manual intervention, and also improves the stability, energy efficiency and safety of the overall production.

Description

Technical Field

[0001] The invention relates to the technical field of aluminum profile extrusion control, and in particular to a high-efficiency energy-consuming extrusion speed regulating system and method. Background Art

[0002] The extruder is the main equipment for the production of light alloy (aluminum alloy, copper alloy and magnesium alloy) tubes, rods and profiles. The extruder is the most important equipment for metal extrusion processing and an important method of forming metal by plastic pressure. Its important feature is that it can process metal ingots into tubes, rods and profiles in an instant. In current industrial production, extrusion has become a widely used key technology. At present, one of the characteristics of modern extruders is to develop in the direction of energy saving, thereby reducing the cost of producing aluminum products, and one of the important measures to achieve the above is energy saving. However, the power of extruders is quite large, so energy saving has naturally become the first choice for enterprises. Enterprises can often reduce the power consumption of aluminum in the extrusion process, thereby reducing the energy consumption of aluminum production and ultimately reducing costs.

[0003] However, the existing extruders have many deficiencies in the actual application of adjusting the extrusion speed when extruding aluminum profiles, especially in complex and dynamic production environments. It is difficult to meet the growing efficiency and precision requirements. For example, the existing technology usually relies on fixed rules or simple feedback control methods to adjust the extrusion speed. The main problem of this method is the lack of predictive ability for abnormal situations that may occur in the future. Moreover, since the system cannot foresee potential production problems in advance, such as load fluctuations, equipment aging or changes in material properties, when problems actually occur, the system often reacts with lag and cannot adjust operating parameters in time, resulting in production interruptions or product quality degradation.

[0004] In addition, the existing extrusion speed adjustment system is often adjusted based on a single variable, such as: simple feedback control based on the current speed or pressure. The limitation of this method is that it cannot comprehensively consider multiple influencing factors and it is difficult to achieve the global optimal control of the system; especially in the face of changing production needs and energy consumption targets, it is difficult for the existing system to dynamically adjust the extrusion speed and pressure, thereby failing to achieve the optimization of energy efficiency. This deficiency not only leads to energy waste, but may also affect production efficiency and product quality; and, the traditional extrusion speed adjustment system often requires operators to frequently intervene and manually adjust system parameters, which not only increases labor intensity, but also has a high risk of operating errors; at the same time, the lag and inaccuracy of manual adjustment can easily lead to instability in the production process, thereby affecting product quality and energy efficiency; especially in complex production environments, it is difficult for operators to make the best judgment in a timely manner, which may cause the equipment to run in a suboptimal state for a long time, and also increase the possibility of safety hazards and system failures.

[0005] Therefore, how to intelligently adjust and control the extrusion production of aluminum profiles on a high-efficiency extrusion machine is a technical problem that technicians currently need to solve. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a high-efficiency energy-consuming extrusion speed regulation system and method, which solves the problems mentioned in the background technology.

[0007] To achieve the above objectives, the present invention provides, in a first aspect, a high-efficiency energy-consuming extrusion speed regulating system, comprising:

[0008] Data acquisition and preprocessing module, feature extraction module, prediction and model update module, dynamic load optimization module and real-time regulation and control module;

[0009] The data acquisition and preprocessing module is used to collect equipment operation parameter information and material status information in real time to form an initial data group DAT, and respectively mark time t and time synchronization processing to obtain a time series data group DATts, and then synchronously preprocess the time series data group DATts to obtain a storage data group DATp;

[0010] The feature extraction module is used to extract statistical features and time features from the stored data group DATp, and screen and generate feature vectors FTVs;

[0011] The prediction and model update module is used to establish a training prediction model based on a machine learning algorithm, input the feature vector FTVs into the training prediction model, and train the output result risk prediction index Yt, and match and predict it with the preset extrusion operation prediction status assessment threshold JZ to generate an extrusion operation status assessment result;

[0012] The dynamic load optimization module triggers the regulation of the current extrusion operation state according to the extrusion operation state evaluation result, outputs the evaluation touch execution result, and performs load prediction analysis based on the evaluation touch execution result to obtain the load prediction index LODpred, and simultaneously obtains the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt according to the load prediction index LODpred;

[0013] The real-time adjustment and control module obtains the real-time extrusion speed SPD and extrusion pressure PRS of the current production area, compares them with the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt, and adjusts and controls the real-time extrusion speed SPD and the extrusion pressure PRS according to the comparison results.

[0014] Preferably, the data acquisition and preprocessing module includes a sensor acquisition unit and a data preprocessing unit;

[0015] The sensor acquisition unit is used to collect the operating parameter information and discharging status information of the extruder equipment in real time, and form an initial data group DAT, and mark the time t and time synchronization processing of the initial data group DAT respectively, to obtain a time series data group DATts, the time series data group DATts={TEM(t), PRS(t), SPD(t), VSC(t), ELM(t), LOD(t)}; wherein, the initial data group DAT includes extrusion temperature TEM, extrusion pressure PRS, extrusion speed SPD, material viscosity VSC, material elastic modulus ELM and production load LOD; the data preprocessing unit is used to preprocess the time series data group DATts and generate a storage data group DATp, the storage data group DATp={TEMp(t), PRSp(t), SPDp(t), VSCp(t), ELMp(t), LODp(t)};

[0016] Among them, the preprocessing includes data cleaning processing and data normalization processing. The data cleaning processing includes using a median filter and a Kalman filter to remove noise in the data; the data normalization processing includes using a min-max normalization processing method to eliminate the dimensional differences between different parameter data.

[0017] Preferably, the feature vector FTVs={Mean(X), Var(t), Std(t), Max(t), Min(t), Sks(t), Kts(t)}; the Mean(X) is the mean feature of the central tendency of the data in the stored data group DATp; the Var(t) is the variance feature of the discreteness in the stored data group DATp; the Std(t) is the standard deviation feature of the discreteness in the stored data group DATp; the Max(t) is the maximum feature of the extreme peak value in the stored data group DATp; the Min(t) is the minimum feature of the extreme valley value in the stored data group DATp; the Sks(t) is the skewness feature of the symmetry of the distribution in the stored data group DATp; the Kts(t) is the kurtosis feature of the steepness of the distribution in the stored data group DATp.

[0018] Preferably, the prediction and model updating module includes a modeling unit, which establishes a training prediction model by using a machine learning algorithm, and the training of the training prediction model includes data partitioning processing;

[0019] The data partitioning process is used to divide the feature vector FTVs into a training set, a validation set and a test set, specifically including dividing 70% of the feature vector FTVs into a training set, dividing 15% of the feature vector FTVs into a validation set, and dividing 15% of the feature vector FTVs into a validation set, forming the input data set FTVsr={FTVtrain, FTVval, FTVtest} of the model;

[0020] Preferably, the risk prediction index Yt is obtained by the following calculation formula:

[0021] ;

[0022] In the formula, α represents the pre-weight value, and log represents the logarithmic function, which is specifically used to smooth the result of feature composition. Represents an exponential decay function, which is specifically used to adjust the nonlinear relationship influence of the feature Xi. Represents the coefficient of adjusting attenuation, indicating Sigmoid function, specifically represents the nonlinear effect of eigenvalue Xi in risk prediction, and the output range is [0, 1]. It represents the sine transformation of the quadratic term of the eigenvalue Xi, which is specifically used to capture the periodic fluctuations between eigenvalues. Xi represents the i-th eigenvalue in the eigenvector FTVsr. represents the sine transform adjustment amplitude coefficient, m represents the total number of nonlinear relationship features in the feature vector FTVsr, n represents the total number of polynomial and exponential decay effects in the feature vector FTVsr, Represents the polynomial coefficients, specifically used to adjust The influence of Xj represents the jth eigenvalue in the eigenvector FTVsr. represents the exponential decay term, It is used to represent the attenuation effect of the eigenvalue Xj and the attenuation coefficient, which is specifically used to control the attenuation rate of the exponential attenuation term. Indicates the adjustment factor of the exponential decay term.

[0023] Preferably, the prediction and model updating module further comprises a state matching unit, which is used to match the preset extrusion operation prediction state assessment threshold JZ with the risk prediction index Yt, and generate an extrusion operation state assessment result of the production area, wherein the performance evaluation of the training prediction model comprises using accuracy and mean square error to evaluate the prediction ability of the training prediction model;

[0024] The evaluation results of the extrusion operation status of the production area are obtained through the following matching methods:

[0025] ;

[0026] When the states return state is 1, it is determined that the extrusion operation state evaluation result of the extruder in the production area is an abnormal result;

[0027] When the states return state is 0, it is determined that the extrusion operation state assessment result of the extruder in the production area is normal; the states represents the extrusion operation state assessment result of the production area, the Yt represents the risk prediction index, and the JZ represents the extrusion operation prediction state assessment threshold.

[0028] Preferably, the dynamic load optimization module includes a load prediction unit and a regulation generation unit;

[0029] The load prediction unit triggers and adjusts the extrusion operation state of the current extruder according to the extrusion operation state evaluation result of the extruder, and touches and executes the extrusion operation state evaluation result of the extruder;

[0030] The trigger regulation is specifically triggered by the following mechanisms:

[0031] When the evaluation result of the extrusion operation status of the production area is abnormal, and the states return state is 1, the extrusion operation status evaluation touch execution result of the extruder is executed;

[0032] When the extrusion operation status evaluation result of the production area is normal and the states return state is 0, the extrusion operation status evaluation touch result of the extruder is not executed.

[0033] Preferably, the control generation unit is used for load prediction analysis to obtain a load prediction index LODpred, and synchronously obtains the control extrusion speed SPDopt and the control extrusion pressure PRSopt according to the load prediction index LODpred; wherein the load prediction index LODpred is obtained by the following calculation formula:

[0034] ;

[0035] In the formula, represents the preset weight value, which is specifically used to adjust the influence of the load prediction index, w represents the total number of feature data included in the feature vector FTVsr, Xk represents the kth feature value in the feature vector FTVsr, represents the preset weight value of the kth eigenvalue, represents the time attenuation coefficient of the kth eigenvalue, t represents the time point t, and Tmax represents the upper limit of the time window length;

[0036] The regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt are obtained by the following calculation formula:

[0037] ;

[0038] ;

[0039] In the formula, A represents the reference extrusion speed, B represents the response coefficient of the extrusion speed to the load prediction, C represents the lower limit of the extrusion speed, D represents the reference extrusion pressure, E represents the upper limit of the extrusion pressure, and F represents the response coefficient of the extrusion pressure to the load prediction.

[0040] Preferably, the real-time adjustment and control module obtains the real-time extrusion speed SPD and extrusion pressure PRS, and compares them with the obtained regulated extrusion speed SPDopt and regulated extrusion pressure PRSopt, respectively, determines the extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t), and adjusts the real-time extrusion speed SPD and the extrusion pressure PRS by using PLC and servo motor control, and the operating parameters of the extruder include correcting the real-time extrusion speed SPD and the extrusion pressure PRS according to the extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t);

[0041] The extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t) are obtained by the following calculation formula:

[0042] ;

[0043] ;

[0044] Wherein, SPDopt(t) and PRSopt(t) represent the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt at time t, respectively; SPD(t) and PRS(t) represent the real-time extrusion speed SPD and the extrusion pressure PRS at time t, respectively.

[0045] In a second aspect, the present application provides a high-efficiency energy consumption extrusion speed regulation method, comprising any one of the high-efficiency energy consumption extrusion speed regulation systems described above, characterized in that it comprises:

[0046] Step 1: The data acquisition and preprocessing module collects the extruder equipment operating parameter information and material status information in real time to form the initial data group DAT, and then performs marking time t and time synchronization processing to obtain the time series data group DATts, and then preprocesses the time series data group DATts synchronously, including data cleaning and data normalization, to form the processed storage data group DATp;

[0047] Step 2: The feature extraction module extracts statistical features and time features from the stored data set DATp, and then applies a feature selection algorithm to filter features to form a feature vector FTVs;

[0048] Step 3: The prediction and model update module uses a machine learning algorithm to establish a training prediction model, takes the acquired feature vector FTVs as input features, and trains the established training prediction model, including data partitioning, parameter optimization and hyperparameter tuning of the training prediction model, obtains the output result risk prediction index Yt of the training prediction model, and matches it with the preset extrusion operation prediction status assessment threshold JZ to obtain the extrusion operation status assessment result of the production area;

[0049] Step 4: The dynamic load optimization module triggers and adjusts the current extrusion operation state according to the extrusion operation state evaluation result of the production area, obtains the extrusion operation state evaluation touch execution result of the extruder, and performs load prediction analysis to obtain the load prediction index LODpred, and simultaneously obtains the control extrusion speed SPDopt and the control extrusion pressure PRSopt according to the load prediction index LODpred;

[0050] Step 5: Real-time adjustment and control The real-time extrusion speed SPD and extrusion pressure PRS obtained in the current production area are compared with the obtained regulated extrusion speed SPDopt and regulated extrusion pressure PRSopt, and the real-time extrusion speed SPD and extrusion pressure PRS are adjusted by using PID control.

[0051] The present invention provides a high-efficiency energy-consuming extrusion speed regulation system and method, which has the following beneficial effects:

[0052] (1) When the system is running, by installing sensor groups at multiple nodes of the production line, the equipment operating parameters and material status information are collected in real time to ensure the accuracy and timeliness of the data. Through in-depth feature extraction and screening of the collected data, the system can build a high-precision prediction model and predict potential abnormal situations in advance through machine learning algorithms. When possible risks are predicted, the system can automatically trigger dynamic load optimization and intelligently adjust the extrusion speed and pressure to achieve the optimization of production efficiency and energy efficiency. In addition, through real-time PID control, the system can quickly respond to changes in production conditions, ensure the stability and efficiency of the extrusion process, significantly reduce dependence on manual intervention, and reduce the risk of operational errors. This highly integrated intelligent control system not only overcomes the shortcomings of the existing technology, such as lack of prediction ability, insufficient dynamic adjustment, delayed response and dependence on manual intervention, but also greatly improves the stability, energy efficiency and safety of the overall production.

[0053] (2) Based on the feature vector FTVs, a high-precision risk prediction index Yt is generated through data partitioning and model training, which can predict potential abnormal situations in advance. At the same time, the state matching unit further combines the preset extrusion operation prediction state evaluation threshold JZ to accurately evaluate the operating status of the production area. This method can not only identify abnormal situations, but also optimize the production process through quantitative prediction results, reducing production interruptions and quality fluctuations caused by sudden problems. The system also ensures the efficiency and accuracy of the prediction model through multi-dimensional model evaluation, reduces the subjectivity of human judgment, and improves the automation and intelligence level of the system.

[0054] (3) According to the evaluation results of the extrusion operation status in the production area, the control mechanism is triggered intelligently to ensure that load adjustments are made only when necessary, thereby avoiding unnecessary system intervention and improving the overall operation efficiency. The control generation unit further generates a load prediction index LODpred through complex load prediction analysis, and dynamically calculates the optimal control extrusion speed SPDopt and control extrusion pressure PRSopt based on the index. This process ensures that the extrusion production equipment always operates in the optimal state, reduces energy consumption and improves production quality. At the same time, the system corrects the deviation of extrusion speed and pressure in real time through precise PID control, making the production process more stable and reliable. This intelligent load prediction and control mechanism not only improves the system's adaptability, but also significantly optimizes resource utilization and production efficiency, ensuring the continuous and efficient operation of the production line in a complex environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic diagram of a block diagram of a high-efficiency energy-consuming extrusion speed regulating system of the present invention;

[0056] Figure 2 The present invention is a schematic diagram of the steps of a high-efficiency energy-consuming extrusion speed adjustment method. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Example 1

[0058] The present invention provides a high-efficiency energy consumption extrusion speed regulation system, please refer to Figure 1 , including data acquisition and preprocessing module, feature extraction module, prediction and model updating module, dynamic load optimization module and real-time regulation and control module;

[0059] The data acquisition and preprocessing module is used to collect equipment operation parameter information and material status information in real time to form an initial data group DAT, and respectively perform marking time t and time synchronization processing to obtain a time series data group DATts, and then synchronously preprocess the time series data group DATts to obtain a storage data group DATp;

[0060] The feature extraction module is used to extract statistical features and time features from the stored data group DATp, and screen and generate feature vectors FTVs;

[0061] The prediction and model updating module is used to establish a training prediction model based on a machine learning algorithm, input the feature vector FTVs into the training prediction model, train the output result risk prediction index Yt, and match and predict it with the preset extrusion operation prediction status assessment threshold JZ to generate the production area extrusion operation status assessment result;

[0062] The dynamic load optimization module triggers and adjusts the current extrusion operation state according to the evaluation result of the extrusion operation state of the production area, and outputs the evaluation touch execution result of the extrusion operation state, and performs load prediction analysis based on the evaluation touch execution result to obtain the load prediction index LODpred, and simultaneously obtains the control extrusion speed SPDopt and the control extrusion pressure PRSopt according to the load prediction index LODpred;

[0063] The real-time adjustment and control module obtains the real-time extrusion speed SPD and extrusion pressure PRS of the current production area, compares them with the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt, and adjusts and controls the real-time extrusion speed SPD and the extrusion pressure PRS according to the comparison results.

[0064] In this embodiment, by installing sensor groups at multiple nodes of the production line, equipment operating parameters and material status information are collected in real time to ensure the accuracy and timeliness of the data. Through in-depth feature extraction and screening of the collected data, the system can build a high-precision prediction model and predict potential abnormal situations in advance through machine learning algorithms. When possible risks are predicted, the system can automatically trigger dynamic load optimization and intelligently adjust the extrusion speed and pressure to achieve the optimization of production efficiency and energy efficiency. In addition, through real-time PID control, the system can quickly respond to changes in production conditions, ensure the stability and efficiency of the extrusion process, significantly reduce dependence on manual intervention, and reduce the risk of operational errors. This highly integrated intelligent control system not only overcomes the shortcomings of the prior art, such as lack of predictive ability, insufficient dynamic adjustment, delayed response, and reliance on manual intervention, but also greatly improves the stability, energy efficiency, and safety of the overall production.

[0065] Example 2

[0066] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the data acquisition and preprocessing module includes a sensor acquisition unit;

[0067] The sensor acquisition unit is used to collect the operating parameter information and discharge status information of the extruder equipment in real time, and form an initial data group DAT, and mark the time t and time synchronization processing of the initial data group DAT respectively to obtain a time series data group DATts, wherein the time series data group DATts={TEM(t), PRS(t), SPD(t), VSC(t), ELM(t), LOD(t)}; wherein the initial data group DAT includes extrusion temperature TEM, extrusion pressure PRS, extrusion speed SPD, material viscosity VSC, material elastic modulus ELM and production load LOD.

[0068] The data acquisition and preprocessing module also includes a data preprocessing unit, which is used to preprocess the time series data group DATts and generate a storage data group DATp, wherein the storage data group DATp={TEMp(t), PRSp(t), SPDp(t), VSCp(t), ELMp(t), LODp(t)};

[0069] Among them, the preprocessing includes data cleaning processing and data normalization processing. The data cleaning processing includes using a median filter and a Kalman filter to remove noise in the data; the data normalization processing includes using a min-max normalization processing method to eliminate the dimensional differences between different parameter data.

[0070] The feature vector FTVs={Mean(X), Var(t), Std(t), Max(t), Min(t), Sks(t), Kts(t)}; Mean(X) is the mean characteristic of the central tendency of the data in the stored data set DATp; Var(t) is the variance characteristic of the discreteness in the stored data set DATp; Std(t) is the standard deviation characteristic of the discreteness in the stored data set DATp; Max(t) is the maximum characteristic of the extreme peak value in the stored data set DATp; Min(t) is the minimum characteristic of the extreme valley value in the stored data set DATp; Sks(t) is the skewness characteristic of the symmetry of the distribution in the stored data set DATp; Kts(t) is the kurtosis characteristic of the steepness of the distribution in the stored data set DATp;

[0071] Temporal feature extraction includes obtaining data trends and interdependencies in the stored data set DATp using autocorrelation functions and moving average functions.

[0072] Example 3

[0073] This embodiment is explained in Example 2. Please refer to Figure 1 , specifically: the prediction and model updating module includes a modeling unit, the modeling unit establishes a training prediction model by using a machine learning algorithm, and the training of the training prediction model includes data partitioning processing;

[0074] The data partitioning process is used to divide the feature vector FTVs into a training set, a validation set and a test set, specifically including dividing 70% of the feature vector FTVs into a training set, dividing 15% of the feature vector FTVs into a validation set, and dividing 15% of the feature vector FTVs into a validation set, forming an input data set FTVsr={FTVtrain, FTVval, FTVtest} of the model;

[0075] The risk prediction index Yt is obtained by the following calculation formula:

[0076] ;

[0077] In the formula, α represents the pre-weight value, and log represents the logarithmic function, which is specifically used to smooth the result of feature composition. Represents an exponential decay function, which is specifically used to adjust the nonlinear relationship influence of the feature Xi. Represents the coefficient of adjusting attenuation, indicating Sigmoid function, specifically represents the nonlinear effect of eigenvalue Xi in risk prediction, and the output range is [0, 1]. It represents the sine transformation of the quadratic term of the eigenvalue Xi, which is specifically used to capture the periodic fluctuations between eigenvalues. Xi represents the i-th eigenvalue in the eigenvector FTVsr. represents the sine transform adjustment amplitude coefficient, m represents the total number of nonlinear relationship features in the feature vector FTVsr, n represents the total number of polynomial and exponential decay effects in the feature vector FTVsr, Represents the polynomial coefficients, specifically used to adjust The influence of Xj represents the jth eigenvalue in the eigenvector FTVsr. represents the exponential decay term, It is used to represent the attenuation effect of the eigenvalue Xj and the attenuation coefficient, which is specifically used to control the attenuation rate of the exponential attenuation term. Indicates the adjustment factor of the exponential decay term.

[0078] The prediction and model updating module also includes a state matching unit, which is used to match the preset extrusion operation prediction state evaluation threshold JZ with the risk prediction index Yt, and generate an evaluation result of the extrusion operation state in the production area, wherein the performance evaluation of the training prediction model includes using accuracy and mean square error to evaluate the prediction ability of the training prediction model;

[0079] The evaluation result of the extrusion operation status of the production area is obtained by the following matching method:

[0080] ;

[0081] When the state returned by states is 1, it is determined that the extrusion operation state evaluation result of the extruder in the production area is an abnormal result;

[0082] When the states return state is 0, it is determined that the extrusion operation status evaluation result of the extruder in the production area is normal.

[0083] In this embodiment, according to the feature vector FTVs, through data partitioning and model training, a high-precision risk prediction index Yt is generated, which can predict potential abnormal situations in advance. At the same time, the state matching unit further combines the preset extrusion operation prediction state evaluation threshold JZ to accurately evaluate the operating status of the production area. This method can not only identify abnormal situations, but also optimize the production process through quantitative prediction results, and reduce production interruptions and quality fluctuations caused by sudden problems. The system also ensures the efficiency and accuracy of the prediction model through multi-dimensional model evaluation, reduces the subjectivity of human judgment, and improves the automation and intelligence level of the system.

[0084] Example 4

[0085] This embodiment is explained in Example 3, please refer to Figure 1 ,Specifically: the dynamic load optimization module includes a load prediction unit and a control generation unit;

[0086] The load prediction unit triggers and adjusts the current extrusion operation state according to the extrusion operation state evaluation result of the extruder equipment in the production area, and obtains the extrusion operation state evaluation touch execution result of the extruder in the production area;

[0087] The triggering and regulating mode is specifically triggered by the following mechanism:

[0088] When the extrusion operation status evaluation result of the production area is abnormal and the states return status is 1, the extrusion operation status evaluation touch execution result of the extruder in the production area is obtained;

[0089] When the extrusion operation status evaluation result of the production area is no abnormality and the states return status is 0, the extrusion operation status evaluation touch non-execution result of the production area is obtained;

[0090] The control generation unit performs load prediction analysis to obtain a load prediction index LODpred, and simultaneously obtains a control extrusion speed SPDopt and a control extrusion pressure PRSopt according to the load prediction index LODpred.

[0091] The load prediction index LODpred is obtained by the following calculation formula:

[0092] ;

[0093] In the formula, represents the preset weight value, which is specifically used to adjust the influence of the load prediction index, w represents the total number of feature data included in the feature vector FTVsr, Xk represents the kth feature value in the feature vector FTVsr, represents the preset weight value of the kth eigenvalue, represents the time attenuation coefficient of the kth eigenvalue, t represents the time point t, and Tmax represents the upper limit of the time window length;

[0094] The regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt are obtained by the following calculation formula:

[0095] ;

[0096] ;

[0097] In the formula, A represents the reference extrusion speed, B represents the response coefficient of the extrusion speed to the load prediction, C represents the lower limit of the extrusion speed, D represents the reference extrusion pressure, E represents the upper limit of the extrusion pressure, and F represents the response coefficient of the extrusion pressure to the load prediction.

[0098] The real-time adjustment and control module compares the real-time extrusion speed SPD and extrusion pressure PRS obtained in the current production area with the obtained regulated extrusion speed SPDopt and regulated extrusion pressure PRSopt, obtains the extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t), and adjusts the operating parameters of the extruder by controlling the real-time extrusion speed SPD and the extrusion pressure PRS by using PLC and servo motor, including correcting the real-time extrusion speed SPD and the extrusion pressure PRS according to the extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t);

[0099] The extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t) are obtained by the following calculation formula:

[0100] ;

[0101] ;

[0102] Wherein, SPDopt(t) and PRSopt(t) represent the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt at time t, respectively; SPD(t) and PRS(t) represent the real-time extrusion speed SPD and the extrusion pressure PRS at time t, respectively.

[0103] Wherein, SPDopt(t) and PRSopt(t) represent the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt at time t, respectively; SPD(t) and PRS(t) represent the real-time extrusion speed SPD and the extrusion pressure PRS at time t, respectively.

[0104] In this embodiment, the control mechanism is triggered intelligently according to the evaluation results of the extrusion operation status of the production area to ensure that load adjustments are made only when necessary, thereby avoiding unnecessary system interventions and improving the overall operation efficiency. The control generation unit further generates a load prediction index LODpred through complex load prediction analysis, and dynamically calculates the optimal control extrusion speed SPDopt and control extrusion pressure PRSopt based on the index. This process ensures that the production equipment always operates in the optimal state, reduces energy consumption and improves production quality. At the same time, the system corrects the deviation of extrusion speed and pressure in real time through precise PID control, making the production process more stable and reliable. This intelligent load prediction and control mechanism not only improves the adaptive ability of the system, but also significantly optimizes resource utilization and production efficiency, ensuring the continuous and efficient operation of the production line in a complex environment.

[0105] Example 5

[0106] Please refer to Figure 2 The present application also provides a high-efficiency energy-consuming extrusion speed adjustment method, which specifically includes the following steps:

[0107] Step 1: The data acquisition and preprocessing module installs sensor groups at multiple production nodes of the production line to collect equipment operation parameter information and material status information in real time to form an initial data group DAT, and then marks the time t and performs time synchronization processing to obtain a time series data group DATts. The time series data group DATts is then preprocessed synchronously, including data cleaning and data normalization, to form a processed storage data group DATp;

[0108] Step 2: The feature extraction module extracts statistical features and time features from the stored data set DATp, and then applies a feature selection algorithm to filter features to form a feature vector FTVs;

[0109] Step 3: The prediction and model update module uses a machine learning algorithm to establish a training prediction model, takes the acquired feature vector FTVs as input features, and trains the established training prediction model, including data partitioning, parameter optimization and hyperparameter tuning of the training prediction model, obtains the output result risk prediction index Yt of the training prediction model, and matches it with the preset extrusion operation prediction status assessment threshold JZ to obtain the extrusion operation status assessment result of the production area;

[0110] Step 4: The dynamic load optimization module triggers and adjusts the current extrusion operation state according to the extrusion operation state evaluation result of the production area, obtains the extrusion operation state evaluation touch execution result of the extruder in the production area, and performs load prediction analysis to obtain the load prediction index LODpred, and simultaneously obtains the control extrusion speed SPDopt and the control extrusion pressure PRSopt according to the load prediction index LODpred;

[0111] Step 5: Real-time adjustment and control The real-time extrusion speed SPD and extrusion pressure PRS obtained in the current production area are compared with the obtained regulated extrusion speed SPDopt and regulated extrusion pressure PRSopt, and the real-time extrusion speed SPD and extrusion pressure PRS are adjusted by using PID control.

[0112] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A high-efficiency energy-consuming extrusion speed regulation system, characterized in that: include: Data acquisition and preprocessing module, feature extraction module, prediction and model update module, dynamic load optimization module and real-time regulation and control module; The data acquisition and preprocessing module is used to collect equipment operation parameter information and material status information in real time to form an initial data group DAT, and respectively perform marking time t and time synchronization processing to obtain a time series data group DATts, and then synchronously preprocess the time series data group DATts to obtain a storage data group DATp. The initial data group includes extrusion temperature TEM, extrusion pressure PRS, extrusion speed SPD, material viscosity VSC, material elastic modulus ELM and production load LOD; The feature extraction module is used to extract statistical features and time features from the stored data group DATp, and screen and generate feature vectors FTVs; The prediction and model updating module is used to establish a training prediction model based on a machine learning algorithm, input the feature vector FTVs into the training prediction model, train the output result risk prediction index Yt, and match and predict it with the preset extrusion operation prediction state assessment threshold JZ to generate an extrusion operation state assessment result; The dynamic load optimization module triggers the regulation of the current extrusion operation state according to the extrusion operation state evaluation result, outputs the evaluation touch execution result, and performs load prediction analysis based on the evaluation touch execution result to obtain the load prediction index LODpred, and simultaneously obtains the regulation of the extrusion speed SPDopt and the regulation of the extrusion pressure PRSopt according to the load prediction index LODpred; The real-time adjustment and control module obtains the real-time extrusion speed SPD and extrusion pressure PRS of the current production area, compares them with the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt, and adjusts and controls the real-time extrusion speed SPD and the extrusion pressure PRS according to the comparison results.

2. According to claim 1, a high-efficiency energy consumption extrusion speed regulation system is characterized in that: The data acquisition and preprocessing module includes a sensor acquisition unit and a data preprocessing unit; The sensor acquisition unit is used to collect the operating parameter information and discharging status information of the extruder equipment in real time, and form an initial data group DAT, and mark the time t and time synchronization processing of the initial data group DAT respectively to obtain a time series data group DATts, the time series data group DATts = {TEM(t), PRS(t), SPD(t), VSC(t), ELM(t), LOD(t)}; wherein the initial data group DAT includes extrusion temperature TEM, extrusion pressure PRS, extrusion speed SPD, material viscosity VSC, material elastic modulus ELM and production load LOD; the data preprocessing unit is used to preprocess the time series data group DATts and generate a storage data group DATp, the storage data group DATp = {TEMp(t), PRSp(t), SPDp(t), VSCp(t), ELMp(t), LODp(t)}; Among them, the preprocessing includes data cleaning processing and data normalization processing. The data cleaning processing includes using a median filter and a Kalman filter to remove noise in the data; the data normalization processing includes using a min-max normalization processing method to eliminate the dimensional differences between different parameter data.

3. The high-efficiency energy-consuming extrusion speed regulation system according to claim 1 is characterized in that: The feature vector FTVs = {Mean(X), Var(t), Std(t), Max(t), Min(t), Sks(t), Kts(t)}; the Mean(X) is the mean characteristic of the central tendency composition of the measured data in the storage data set DATp; The Var(t) is the variance characteristic of the discreteness component in the stored data set DATp; the Std(t) is the standard deviation characteristic of the discreteness component in the stored data set DATp; The Max(t) is the maximum value characteristic of the extreme peak value composition in the stored data set DATp; the Min(t) is the minimum value characteristic of the extreme valley value composition in the stored data set DATp; The Sks(t) is the skewness characteristic of the symmetry of the distribution in the storage data set DATp; the Kts(t) is the kurtosis characteristic of the steepness of the distribution in the storage data set DATp.

4. The high-efficiency energy-consuming extrusion speed regulation system according to claim 1 is characterized in that: The prediction and model updating module includes a modeling unit, wherein the modeling unit establishes a training prediction model by using a machine learning algorithm, and the training of the training prediction model includes data partitioning processing; The data partitioning process is used to divide the feature vector FTVs into a training set, a validation set and a test set, specifically including dividing 70% of the feature vector FTVs into a training set, 15% of the feature vector FTVs into a validation set, and 15% of the feature vector FTVs into a test set, forming an input data set FTVsr = {FTVtrain, FTVval, FTVtest} of the model; 5. The high-efficiency energy-consuming extrusion speed regulation system according to claim 1 is characterized in that: The risk prediction index Yt is obtained by the following calculation formula: In the formula, α represents the pre-weight value, log represents the logarithmic function, which is specifically used to smooth the result of feature composition, 1+e -βi·Xi Represents an exponential decay function, which is specifically used to adjust the nonlinear relationship influence of the feature Xi, βi represents the coefficient of adjusting the decay, Represents the Sigmoid function, specifically the nonlinear effect of the eigenvalue Xi in risk prediction, with an output range of [0, 1], sin(γi·Xi 2 )·Xi represents the sine transform of the quadratic term of the eigenvalue Xi, which is specifically used to capture the periodic fluctuations between the eigenvalues. Xi represents the i-th eigenvalue in the eigenvector FTVsr. γi represents the sine transform adjustment amplitude coefficient. m represents the total number of nonlinear relationship features in the eigenvector FTVsr. n represents the total number of polynomial and exponential decay effects in the eigenvector FTVsr. Represents the polynomial coefficients, which are used to adjust Xj 3 The influence of Xj is the jth eigenvalue in the eigenvector FTVsr, e -υj·Xj represents the exponential decay term, which is used to represent the decay effect of the eigenvalue Xj. εj represents the decay coefficient, which is specifically used to regulate the decay rate of the exponential decay term. υj represents the adjustment coefficient of the exponential decay term.

6. The high-efficiency energy-consuming extrusion speed regulation system according to claim 1 is characterized in that: The prediction and model updating module also includes a state matching unit, which is used to match the preset extrusion operation prediction state evaluation threshold JZ with the risk prediction index Yt, and generate an evaluation result of the extrusion operation state in the production area, wherein the performance evaluation of the training prediction model includes using accuracy and mean square error to evaluate the prediction ability of the training prediction model; The evaluation result of the extrusion operation status of the production area is obtained by the following matching method: When the states return state is 1, it is determined that the extrusion operation state evaluation result of the extruder in the production area is an abnormal result; When the states return state is 0, it is determined that the extrusion operation state assessment result of the extruder in the production area is normal; the states represents the extrusion operation state assessment result of the production area, the Yt represents the risk prediction index, and the JZ represents the extrusion operation prediction state assessment threshold.

7. The high-efficiency energy-consuming extrusion speed regulation system according to claim 6 is characterized in that: The dynamic load optimization module includes a load prediction unit and a regulation generation unit; The load prediction unit triggers and adjusts the extrusion operation state of the current extruder according to the extrusion operation state evaluation result of the extruder, and touches and executes the extrusion operation state evaluation result of the extruder; The triggering regulation is specifically triggered by the following mechanisms: When the evaluation result of the extrusion operation state of the production area is abnormal, and the states return state is 1, the extrusion operation state evaluation touch execution result of the extruder is executed; When the evaluation result of the extrusion operation state of the production area is normal and the states return state is 0, the extrusion operation state evaluation touch result of the extruder is not executed.

8. The high-efficiency energy-consuming extrusion speed regulation system according to claim 7 is characterized in that: The control generation unit is used for load prediction analysis to obtain a load prediction index LODpred, and synchronously obtains a control extrusion speed SPDopt and a control extrusion pressure PRSopt according to the load prediction index LODpred; wherein the load prediction index LODpred is obtained by the following calculation formula: In the formula, represents a preset weight value, which is specifically used to adjust the influence of the load prediction index, w represents the total number of feature data included in the feature vector FTVsr, Xk represents the kth eigenvalue in the feature vector FTVsr, θk represents the preset weight value of the kth eigenvalue, τk represents the time attenuation coefficient of the kth eigenvalue, t represents the time point t, and Tmax represents the upper limit of the time window length; The regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt are obtained by the following calculation formula: PRSopt = D·(EF·LODpred); In the formula, A represents the reference extrusion speed, B represents the response coefficient of the extrusion speed to the load prediction, C represents the lower limit of the extrusion speed, D represents the reference extrusion pressure, E represents the upper limit of the extrusion pressure, and F represents the response coefficient of the extrusion pressure to the load prediction.

9. The high-efficiency energy-consuming extrusion speed regulation system according to claim 1 is characterized in that: The real-time adjustment and control module obtains the real-time extrusion speed SPD and extrusion pressure PRS, and compares them with the obtained regulated extrusion speed SPDopt and regulated extrusion pressure PRSopt, respectively, to determine the extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t), and adjusts the real-time extrusion speed SPD and the extrusion pressure PRS by using PLC and servo motor control, and the operating parameters of the extruder include correcting the real-time extrusion speed SPD and the extrusion pressure PRS according to the extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t); The extrusion speed error value Espd(t) and the extrusion pressure error value Eprs(t) are obtained by the following calculation formula: Espd(t)=SPDopt(t)-SPD(t); Eprs(t)=PRSopt(t)-PRS(t); Wherein, SPDopt(t) and PRSopt(t) represent the regulated extrusion speed SPDopt and the regulated extrusion pressure PRSopt at time t, respectively; SPD(t) and PRS(t) represent the real-time extrusion speed SPD and the extrusion pressure PRS at time t, respectively.

10. A high-efficiency energy consumption extrusion speed regulation method, comprising a high-efficiency energy consumption extrusion speed regulation system as claimed in any one of claims 1 to 9, characterized in that: include: Step 1: The data acquisition and preprocessing module collects the extruder equipment operating parameter information and material status information in real time to form an initial data group DAT, and then performs marking time t and time synchronization processing to obtain a time series data group DATts, and then preprocesses the time series data group DATts synchronously, including data cleaning and data normalization, to form a processed storage data group DATp, wherein the initial data group includes extrusion temperature TEM, extrusion pressure PRS, extrusion speed SPD, material viscosity VSC, material elastic modulus ELM and production load LOD; Step 2: The feature extraction module extracts statistical features and time features from the stored data set DATp, and then applies a feature selection algorithm to filter features to form a feature vector FTVs; Step 3: The prediction and model update module uses a machine learning algorithm to establish a training prediction model, takes the acquired feature vector FTVs as input features, and trains the established training prediction model, including data partitioning, parameter optimization and hyperparameter tuning of the training prediction model, obtains the output result risk prediction index Yt of the training prediction model, and matches it with the preset extrusion operation prediction status assessment threshold JZ to obtain the extrusion operation status assessment result of the production area; Step 4: The dynamic load optimization module triggers and adjusts the current extrusion operation state according to the extrusion operation state evaluation result of the production area, obtains the extrusion operation state evaluation touch execution result of the extruder, and performs load prediction analysis to obtain the load prediction index LODpred, and simultaneously obtains the control extrusion speed SPDopt and the control extrusion pressure PRSopt according to the load prediction index LODpred; Step 5: Real-time adjustment and control The real-time extrusion speed SPD and extrusion pressure PRS obtained in the current production area are compared with the obtained regulated extrusion speed SPDopt and regulated extrusion pressure PRSopt, and the real-time extrusion speed SPD and extrusion pressure PRS are adjusted by using PID control.

Images