Intelligent control method and system for aluminum alloy profile production

Abstract

The invention relates to the technical field of aluminum alloy profile production, and discloses an intelligent control method and system for aluminum alloy profile production, and the method comprises the steps: collecting vibration frequency and amplitude data, and extracting periodic features, abnormal peak features and dominant frequency offset through Fourier transform; when the dominant frequency offset exceeds an offset threshold value, the vibration abnormal degree is evaluated, and an abnormal offset evaluation result is obtained; inputting the vibration signal features into a pre-established health state classifier to obtain a health state label; when the health state label is displayed to be abnormal, generating initial operation parameter configuration according to an abnormal state mapping table; according to the initial operation parameter configuration, updating an equipment operation state through a real-time control system to obtain new operation data; and extracting a new dominant frequency offset according to the new operation data, when the new dominant frequency offset exceeds an offset threshold, updating the abnormal state mapping table, and generating optimized operation parameter configuration. The method can improve the reliability and efficiency of equipment operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aluminum alloy profile production, and particularly relates to an intelligent control method and system for aluminum alloy profile production. BACKGROUND

[0002] At present, in the manufacturing industry, the production of high-quality aluminum alloy plates is a core link supporting high-end industries such as construction, automobiles and aviation, and the performance and quality thereof directly determine the reliability and competitiveness of terminal products. At present, with the increasing demand for high-quality aluminum alloy plates in the market, intelligent control in the production process is particularly important. However, due to the complexity of the aluminum alloy rolling process, there is a highly nonlinear relationship between the equipment operating parameters and the vibration characteristics, which makes it difficult to achieve precise regulation and control simply by relying on manual experience. Therefore, how to improve the sensing ability and response speed of the equipment state through intelligent signal analysis and parameter dynamic adjustment has become a key technical challenge to realize efficient and stable production of high-quality aluminum alloy plates.

[0003] In one prior art, first, a vibration sensor installed on the bearing seat of the rolling mill is used to collect time-domain vibration signals during equipment operation. Then, after eliminating high-frequency noise interference through sliding average filtering processing, the time-domain signals are converted into frequency-domain spectrograms by using fast Fourier transform (FFT). The operator regularly observes the amplitude changes of the fundamental frequency (corresponding to the main shaft speed) and its harmonic components in the spectrogram, and adjusts the roll reduction amount compensation coefficient according to experience. After each parameter modification, the production needs to be stabilized for 15-20 minutes before new vibration data is collected for effect verification. However, the method of judging vibration characteristics by relying on manual experience in the prior art lacks quantitative standards, and it is difficult to capture early signs of failure in a timely manner. The fixed parameter processing method cannot dynamically adapt to changes in the equipment state, making it difficult to effectively identify potential abnormalities. The discrete manual adjustment mode disrupts the synergy of process parameters, which may even exacerbate equipment vibration. The long verification period makes it impossible to correct abnormal conditions in a timely manner, resulting in continuous accumulation of failure risks. These defects result in long-term operation of the equipment in a non-optimal condition, which not only accelerates the wear of components, but also causes chain-like failures.

[0004] Therefore, the prior art has the problem of low equipment operation reliability and efficiency. SUMMARY

[0005] The present application provides an intelligent control method and system for aluminum alloy profile production to improve the reliability and efficiency of equipment operation.

[0006] In a first aspect, to solve the above technical problems, the present application provides an intelligent control method for aluminum alloy profile production, comprising: collecting operation data, wherein the operation data includes vibration frequency and amplitude data; extract initial vibration signal features based on the vibration frequency and the amplitude data, wherein the initial vibration signal features include periodicity features, abnormal peak features and a main frequency offset; when the main frequency offset exceeds a preset offset threshold, analyze and determine a vibration abnormality degree of the key component to obtain an abnormal offset evaluation result; input the abnormal offset evaluation result and the initial vibration signal features into a pre-established health state classifier to obtain a health state label; when the health state label shows abnormality, generate an initial operation parameter configuration through a pre-established abnormal state mapping table; update the equipment operation state through a real-time control system according to the initial operation parameter configuration to obtain new operation data; extract a new main frequency offset according to the new operation data, and update the abnormal state mapping table to generate an optimized operation parameter configuration when the new main frequency offset exceeds the preset offset threshold.

[0007] In an alternative embodiment, the initial vibration signal features are extracted based on the vibration frequency and the amplitude data through a Fourier transform algorithm, including: perform frequency spectrum decomposition on the vibration frequency and the amplitude data through a Fourier transform algorithm to obtain a frequency domain feature set; identify and extract the amplitude maximum frequency as the main frequency according to the frequency domain feature set, and calculate the offset of the main frequency from a preset base frequency to obtain the main frequency offset; perform time domain analysis on the frequency domain feature set using a sliding window technique according to the frequency domain feature set to obtain periodicity features; when the periodicity features include abnormal points where the amplitude data exceeds a preset amplitude threshold, perform local feature decomposition on the abnormal points through a discrete wavelet transform to determine abnormal peak features; when the periodicity features do not include abnormal points where the amplitude data exceeds the preset amplitude threshold, there are no abnormal peak features; use the main frequency offset, the periodicity features and the abnormal peak features as the initial vibration signal features.

[0008] In an alternative embodiment, when the main frequency offset exceeds the preset offset threshold, the vibration abnormality degree of the key component is analyzed and determined to obtain the abnormal offset evaluation result, including: when the main frequency offset exceeds the preset offset threshold, calculate the similarity score of the periodicity features and pre-stored historical normal data using a cosine similarity algorithm, and use it as the vibration abnormality degree; The vibration anomaly degree and the main frequency offset are weighted and fused to generate an anomaly score; The anomaly score is smoothed by using a mean filtering algorithm to output an anomaly offset evaluation result.

[0009] In an optional embodiment, the construction process of the health state classifier includes: Obtaining historical vibration signal features and corresponding historical health states; An initial health state classifier is constructed by using a support vector machine algorithm, the classification boundary is initialized by using an RBF kernel function, and a penalty coefficient is set to determine an error tolerance threshold; Based on the historical vibration signal features, a kernel density estimation method is used to calculate a curvature adjustment parameter of the classification boundary, and the curvature of the classification boundary is dynamically adjusted by using an RBF kernel function; According to the historical health states, the penalty coefficient is optimized by using a grid search method to determine the error tolerance threshold; The adjusted classification boundary is cross-validated to calculate the classification accuracy after boundary adjustment and the error distribution set; When the error distribution set exceeds the error tolerance threshold, the curvature parameter of the classification boundary is re-optimized by using a gradient descent method, and the classification boundary is iteratively updated until the error distribution set does not exceed the error tolerance threshold; The optimized classification boundary and the health state are matched in a multi-dimensional space by using an RBF kernel function mapping to determine the hyperplane equation corresponding to each health state; Combined with the time sequence characteristics of the historical health states, a sliding window method is used to extract dynamic trend parameters of health state changes, the dynamic trend parameters are associated with the hyperplane equation for associated analysis, and a final health state classifier is constructed.

[0010] In an optional embodiment, when the health state label is displayed as abnormal, an initial operating parameter configuration is generated by using a pre-established abnormal state mapping table, including: When the health state label is displayed as abnormal, the load current of the device is obtained, and the correlation coefficient between the load current and the main frequency offset is calculated by using a Pearson correlation coefficient algorithm to obtain a load correlation coefficient; When the load correlation coefficient exceeds a preset load correlation threshold, the pre-established abnormal state mapping table is queried to generate an initial operating parameter configuration; When the load correlation coefficient does not exceed the preset load correlation threshold, the original operating parameter configuration is maintained to obtain the initial operating parameter configuration.

[0011] In an optional embodiment, the updating of the device operating state by a real-time control system according to the initial operating parameter configuration to obtain new operating data includes: According to the initial operating parameter configuration, the corresponding adjustable parameter range is obtained from a pre-established operating parameter database to determine an adapted adjustment amplitude range; According to the initial operating parameter configuration and the adjustment amplitude range, real-time control instructions are constructed through the industrial transmission protocol to update the equipment operating status and obtain new operating data.

[0012] In an optional embodiment, extracting a new main frequency offset according to the new operating data, and updating the abnormal state mapping table when the new main frequency offset exceeds a preset offset threshold, to generate an optimized operating parameter configuration, includes: Performing spectrum analysis on the new operating data using a Fourier transform algorithm to obtain a new main frequency offset; When the new main frequency offset exceeds a preset offset threshold, the abnormal state mapping table is optimized using a reinforcement learning algorithm in combination with new operating data to generate an optimized operating parameter configuration.

[0013] In a second aspect, the present invention provides an intelligent control system for aluminum alloy profile production, comprising: A data acquisition module, configured to acquire operating data, wherein the operating data includes vibration frequency and amplitude data; a feature analysis module, configured to extract initial vibration signal features using a Fourier transform algorithm based on the vibration frequency and the amplitude data, wherein the initial vibration signal features include periodic features, abnormal peak features, and a main frequency offset; an abnormality analysis module, configured to analyze and determine the degree of vibration abnormality of key components when the main frequency offset exceeds a preset offset threshold, and obtain an abnormal offset assessment result; a health status classification module, configured to input the abnormal deviation assessment result and the initial vibration signal feature into a pre-established health status classifier to obtain a health status label; An operating parameter generation module is used to generate an initial operating parameter configuration through a pre-established abnormal state mapping table when the health status label is displayed as abnormal; A control operation module is used to update the equipment operation status through a real-time control system according to the initial operation parameter configuration to obtain new operation data; An operating parameter optimization module is configured to extract a new main frequency offset based on the new operating data, and when the new main frequency offset exceeds a preset offset threshold, update the abnormal state mapping table to generate an optimized operating parameter configuration. In a third aspect, the present invention further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements any of the aforementioned intelligent control methods for aluminum alloy profile production.

[0014] In a fourth aspect, the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned intelligent control methods for aluminum alloy profile production.

[0015] Compared with the prior art, the present invention has the following beneficial effects: (1) Through the multi-level signal analysis technology of Fourier transform combined with discrete wavelet transform, the frequency domain and time domain features of the equipment vibration signal are extracted simultaneously, which solves the problem that the traditional single analysis method is difficult to fully capture the abnormal characteristics of the equipment, and significantly improves the accuracy and timeliness of fault detection.

[0016] (2) The dynamic loss level classifier constructed using the support vector machine algorithm adaptively adjusts the classification boundary through the RBF kernel function, overcomes the subjectivity and lag defects of manual experience judgment, and realizes intelligent assessment of equipment health status and early fault warning.

[0017] (3) Based on the dynamic mapping mechanism between abnormal status and operating parameters, in conjunction with the real-time closed-loop control system, intelligent optimization and adjustment of equipment operating parameters are achieved, solving the problem of poor system coordination caused by the fixed parameter mode and significantly improving the stability and efficiency of equipment operation.

[0018] (4) By constructing an anomaly assessment model using the cosine similarity algorithm and weighted fusion technology, combined with mean filtering processing, a stable and reliable equipment status evaluation system was established, which provided an accurate basis for preventive maintenance decisions and effectively extended the service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of an intelligent control method for aluminum alloy profile production provided by the first embodiment of the present invention; Figure 2 This is a schematic structural diagram of an intelligent control system for aluminum alloy profile production provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0020] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0021] With reference to Figure 1 The first embodiment of the present application provides an intelligent control method for aluminum alloy profile production, comprising the following steps: S11, collecting operation data, wherein the operation data includes vibration frequency and amplitude data; S12, based on the vibration frequency and the amplitude data, extracting initial vibration signal features through a Fourier transform algorithm, wherein the initial vibration signal features include periodic features, abnormal peak value features and main frequency offset; S13, when the main frequency offset exceeds a preset offset threshold, analyzing and determining the vibration abnormality degree of the key components to obtain an abnormal offset evaluation result; S14, inputting the abnormal offset evaluation result and the initial vibration signal features into a pre-established health state classifier to obtain a health state label; S15, when the health state label shows abnormality, generating initial operation parameter configurations through a pre-established abnormal state mapping table; S16, updating the equipment operation state through a real-time control system according to the initial operation parameter configurations to obtain new operation data; S17, extracting a new main frequency offset according to the new operation data, and when the new main frequency offset exceeds a preset offset threshold, updating the abnormal state mapping table to generate optimized operation parameter configurations.

[0022] In step S11, operation data is collected, wherein the operation data includes vibration frequency and amplitude data.

[0023] Specifically, the vibration signal is collected in real time by a piezoelectric acceleration sensor installed on a key component of the aluminum alloy profile production equipment. The sensor records the three-axis vibration waveform during equipment operation at a sampling frequency of no less than 1000 Hz. The collected raw data contains time series vibration acceleration values, which are converted into standardized vibration frequency and amplitude data after impedance matching and noise suppression by a signal conditioning circuit. The preset sampling frequency threshold is determined according to the Nyquist sampling theorem of the equipment rated speed, ensuring complete capture of vibration characteristics corresponding to the highest working speed. The vibration frequency and amplitude data obtained in this step provide the basis for subsequent feature extraction, and their accuracy directly affects the reliability of anomaly detection. The high-precision sensor and anti-interference design ensure data quality, meeting the stringent requirements of aluminum alloy profile production for equipment condition monitoring.

[0024] In step S12, based on the vibration frequency and the amplitude data, initial vibration signal features are extracted by a Fourier transform algorithm, wherein the initial vibration signal features include periodicity features, abnormal peak value features, and a main frequency offset.

[0025] In a specific embodiment, the initial vibration signal features are extracted based on the vibration frequency and the amplitude data by a Fourier transform algorithm, including: According to the vibration frequency and the amplitude data, frequency spectrum decomposition is performed by a Fourier transform algorithm to obtain a frequency domain feature set; According to the frequency domain feature set, the maximum amplitude frequency is identified and extracted as the main frequency, and the offset of the main frequency from the preset base frequency is calculated to obtain the main frequency offset; According to the frequency domain feature set, a sliding window technique is used to perform time domain analysis on the frequency domain feature set to obtain periodicity features; When the periodicity features contain abnormal points with amplitude data exceeding a preset amplitude threshold, local feature decomposition of the abnormal points is performed by a discrete wavelet transform to determine abnormal peak value features; When the periodicity features do not contain abnormal points with amplitude data exceeding a preset amplitude threshold, there are no abnormal peak value features; The main frequency offset, the periodicity features, and the abnormal peak value features are used as initial vibration signal features.

[0026] Specifically, the time-domain vibration signal (containing time series of vibration amplitude data) obtained in step S11 is taken as input, and first, a fast Fourier transform algorithm is performed on the time-domain vibration signal to convert the time-domain vibration signal into a frequency-domain feature set containing fundamental frequency components and harmonic components, which records the specific frequency value and corresponding amplitude data of each frequency component. Then, the frequency component with the maximum amplitude is identified from the frequency-domain feature set as the main frequency, and the absolute difference between the main frequency and the preset fundamental frequency calibrated by the equipment technical manual is calculated to output the main frequency deviation. Then, the original time-domain vibration signal is processed using a sliding window with a width that is an integer multiple of the device rotation period, and an autocorrelation function analysis is performed on the time-domain waveform data in the window to output a periodicity feature representing the stability of the vibration law, including the adjacent cycle amplitude correlation coefficient and the cycle length standard deviation index. When the time-domain waveform amplitude data in the sliding window exceeds a preset amplitude threshold value determined according to the 95th percentile of the statistical distribution of historical normal operation data of the device, the time-domain waveform segment in the window is intercepted at the time point where the amplitude exceeds the threshold value, and a three-layer multi-scale decomposition is performed through discrete wavelet transform to extract the frequency position and mutation strength of the energy mutation in the local waveform, and output the abnormal peak feature; if there is no amplitude exceeding point in the window, the abnormal peak feature is marked as null. Finally, the main frequency deviation, the adjacent cycle amplitude correlation coefficient and the cycle length standard deviation index in the periodicity feature, and the frequency position and mutation strength in the abnormal peak feature are merged into the initial vibration signal feature. This step integrates time-domain sliding analysis and frequency-domain feature extraction to completely capture the periodicity and transient anomaly of the vibration signal, and provides a multi-dimensional quantitative feature set containing the main frequency deviation, the cycle correlation coefficient, the cycle length standard deviation, the abnormal frequency position, and the abnormal mutation strength for the health state classifier.

[0027] In step S13, when the main frequency deviation exceeds a preset deviation threshold, the vibration abnormality degree of the key component is analyzed and determined to obtain an abnormal deviation evaluation result.

[0028] In a specific embodiment, when the main frequency deviation exceeds a preset deviation threshold, the vibration abnormality degree of the key component is analyzed and determined to obtain an abnormal deviation evaluation result, including: When the main frequency deviation exceeds a preset deviation threshold, a cosine similarity algorithm is used to calculate the similarity score of the periodicity feature with the pre-stored historical normal data, and the similarity score is taken as the vibration abnormality degree; The vibration abnormality degree and the main frequency deviation are weighted and fused to generate an abnormal score; A mean filter algorithm is used to smooth the abnormal score to output the abnormal deviation evaluation result.

[0029] Specifically, the main frequency offset and the periodic features (including the amplitude fluctuation standard deviation and the frequency repeatability) output in step S12 are taken as inputs. When the main frequency offset exceeds a preset offset threshold (determined according to the 99th percentile of the statistical distribution of the main frequency offset during the historical normal operation of the equipment), the amplitude fluctuation standard deviation and the frequency repeatability in the current periodic features are extracted, and a vector is formed with the standard reference values of the corresponding indicators in the pre-stored historical normal data database. The cosine similarity algorithm is used to calculate the direction cosine value of the current feature vector and the standard reference vector, and a similarity score in the range of 0 to 1 is output. The complement of the similarity score is taken as the vibration abnormality degree.

[0030] Then, the vibration abnormality degree and the main frequency offset are weighted and fused according to preset weight coefficients (the weight coefficient of the main frequency offset is 0.6, and the weight coefficient of the vibration abnormality degree is 0.4. The weight distribution is determined based on the historical data analysis of the sensitivity of the feature parameters to faults), to generate an abnormality score in the range of 0 to 100.

[0031] Finally, a mean filtering algorithm with a window width of 5 sampling points is used to smooth the continuously generated abnormality score sequence, to eliminate transient interference fluctuations and output stable abnormality offset evaluation result values. This step quantifies the deviation of the vibration features from the historical normal state, solves the subjectivity problem of manual experience judgment, and provides an objective abnormality degree evaluation basis for health state classification.

[0032] In step S14, the abnormality offset evaluation result and the initial vibration signal features are input into a pre-established health state classifier to obtain a health state label.

[0033] In a specific embodiment, the construction process of the health state classifier includes: acquiring historical vibration signal features and corresponding historical health states; using a support vector machine algorithm to construct an initial health state classifier, initializing a classification boundary through an RBF kernel function, and setting a penalty coefficient to determine a error tolerance threshold; based on the historical vibration signal features, using a kernel density estimation method to calculate a curvature adjustment parameter of the classification boundary, and dynamically adjusting the curvature of the classification boundary through an RBF kernel function; according to the historical health states, optimizing the penalty coefficient through a grid search method to determine the error tolerance threshold; cross-validating the adjusted classification boundary, and calculating the classification accuracy and the error distribution set after the boundary adjustment; when the error distribution set exceeds the error tolerance threshold, re-optimizing the curvature parameter of the classification boundary through a gradient descent method, and iteratively updating the classification boundary until the error distribution set does not exceed the error tolerance threshold; The optimized classification boundary and health status are matched in multi-dimensional space through RBF kernel function mapping to determine the hyperplane equation corresponding to each health status; Combined with the time series characteristics of the historical health status, the sliding window method is used to extract the dynamic trend parameters of the health status change, and the dynamic trend parameters are associated with the hyperplane equation to construct the final health status classifier.

[0034] Specifically, the health status classifier is constructed as follows: First, a dataset of vibration signal features recorded during historical equipment operation is collected. This dataset includes the main frequency offset, the standard deviation of amplitude fluctuations and frequency repetition in periodic features, the frequency location and mutation intensity of abnormal peak features, and the corresponding equipment health status labels (categorized as normal, warning, and abnormal). An initial classification model is constructed using a support vector machine algorithm, with radial basis functions (RBFs) used as the kernel function to initialize the classification boundaries. An initial penalty coefficient of 10 is determined based on the frequency statistics of faults in equipment maintenance records, and an error tolerance threshold of 0.05 is set. Based on the distribution of historical vibration feature data, a kernel density estimation algorithm is used to calculate the density distribution of each health status sample in the feature space. Based on this information, the curvature parameter of the RBF kernel function is dynamically adjusted to ensure that the classification boundaries better match the actual data distribution. The curvature parameter is iteratively optimized using a gradient descent method: Each iteration, the classification error gradient on the validation set is calculated, and the curvature parameter is updated along the negative gradient direction until the classification error rate drops below the preset error tolerance threshold of 0.05. The optimized kernel function is then fixed, and the grid search method is used to search for the optimal penalty coefficient in the range of 0.1 to 100 with a logarithmic step size. Combined with five-fold cross-validation, the parameter combination with the highest classification accuracy is selected. Finally, a hyperplane equation is established to distinguish different health states. At the same time, the historical health state label time series is analyzed, and a sliding window with a window width of 30 sampling points is used to extract the state transition feature matrix, which contains the mean state duration, the probability of adjacent state transitions, and the gradient value of state change. This feature matrix is ​​used as an independent input dimension to fuse with the original vibration features, and an extended feature vector is constructed and input into the optimized support vector machine classifier to form the final health state classification model. The classifier receives a feature vector consisting of the main frequency offset of the current cycle, periodic characteristics, abnormal peak characteristics, and abnormal offset evaluation results, and outputs the device health state label.

[0035] In step S15, when the health status label is displayed as abnormal, an initial operating parameter configuration is generated through a pre-established abnormal status mapping table.

[0036] In a specific embodiment, when the health status label is displayed as abnormal, generating an initial operating parameter configuration through a pre-established abnormal status mapping table includes: When the health status label is displayed as abnormal, the load current of the device is obtained, and a correlation coefficient of the load current and the main frequency offset is calculated by a Pearson correlation coefficient algorithm to obtain a load correlation coefficient; When the load correlation coefficient exceeds a preset load correlation threshold, an initial operation parameter configuration is generated by querying a pre-established abnormal state mapping table. When the load correlation coefficient does not exceed the preset load correlation threshold, the original operation parameter configuration is maintained to obtain the initial operation parameter configuration.

[0037] Specifically, first, the load current data of the current device is obtained, which is derived from the current effective value collected by the current sensor installed in the power loop of the device in real time. The load current data and the main frequency offset obtained in step S12 form a data pair sequence, and the linear correlation of the two is calculated by using the Pearson correlation coefficient algorithm. The algorithm calculates the ratio of the covariance of the two variables to the standard deviation of each variable, and outputs the load correlation coefficient in the range of-1 to 1. The preset load correlation threshold is set to 0.7, which is determined based on the strong correlation between the load and the vibration anomaly in historical data analysis. When the calculated load correlation coefficient exceeds the threshold, the system queries the abnormal state mapping table, which stores the parameter adjustment strategies corresponding to different abnormal types, including the key parameters such as the main shaft speed adjustment amount and the feed rate compensation coefficient. According to the size of the main frequency offset detected at present and the abnormal peak characteristic type, the closest adjustment scheme is matched to generate an initial operation parameter configuration containing specific parameter adjustment values. If the load correlation coefficient does not exceed the threshold, the current operation parameters are maintained unchanged, and the original parameter configuration is output as the initial operation parameter configuration. This step quantifies the correlation between the load and the vibration, ensures the pertinence of parameter adjustment, avoids unnecessary disturbance of the device state, and provides a reasonable initial adjustment scheme for subsequent closed-loop control.

[0038] In step S16, according to the initial operation parameter configuration, the device operation state is updated through a real-time control system to obtain new operation data.

[0039] In a specific embodiment, according to the initial operation parameter configuration, the device operation state is updated through a real-time control system to obtain new operation data, including: According to the initial operation parameter configuration, the corresponding adjustable parameter interval is obtained from the pre-established operation parameter database to determine the adaptive adjustment amplitude interval; According to the initial operation parameter configuration and the adjustment amplitude interval, a real-time control instruction is constructed through an industrial transmission protocol to update the device operation state and obtain new operation data.

[0040] Specifically, first, the adjustable parameter interval matching the current device model and process requirements is retrieved from the operating parameter database, which stores the safe operating range verified by the device manufacturer and the constraint relationship between parameters. For the spindle speed adjustment value and feed rate compensation coefficient proposed in the initial operating parameter configuration, the system automatically matches the corresponding allowed adjustment amplitude interval in the database, the upper and lower limits of which are determined according to the mechanical structure strength of the device and the process quality requirements. Subsequently, the control system compares the adjustment value in the initial operating parameter configuration with the allowed amplitude interval, and if the adjustment value exceeds the interval range, it is automatically truncated to the nearest boundary value to ensure the safety of parameter adjustment. Through the Modbus RTU industrial communication protocol, a control instruction frame containing the device address, function code and adjustment parameter value is constructed, which is transmitted to the device controller through industrial Ethernet to perform parameter update. After the parameter update is completed, the system immediately starts a new round of data acquisition, and obtains the updated vibration frequency and amplitude data as new operating data through the same sensor configuration in step S11, while recording the current timestamp and device working condition information to form a complete operating state snapshot. This step ensures the safety and reliability of device state adjustment through strict parameter interval verification and standardized industrial communication protocol, and provides a new basis for monitoring data for subsequent optimization iteration.

[0041] In step S17, a new main frequency offset is extracted according to the new operating data, and when the new main frequency offset exceeds a preset offset threshold, the abnormal state mapping table is updated to generate an optimized operating parameter configuration.

[0042] In a specific embodiment, the new main frequency offset is extracted according to the new operating data, and when the new main frequency offset exceeds a preset offset threshold, the abnormal state mapping table is updated to generate an optimized operating parameter configuration, including: According to the new operating data, a frequency spectrum analysis is performed through a Fourier transform algorithm to obtain a new main frequency offset; When the new main frequency offset exceeds a preset offset threshold, the abnormal state mapping table is optimized using a reinforcement learning algorithm in combination with the new operating data to generate an optimized operating parameter configuration.

[0043] Specifically, first, the same fast Fourier transform processing as step S12 is performed on the vibration frequency sequence in the new operation data obtained in step S16, the new main frequency component under the current equipment operation state is extracted, and the difference between the component and the equipment rated fundamental frequency is calculated as a new main frequency offset. The preset offset threshold value is the same as the value determined in step S13, that is, the 99th percentile value of the statistical distribution of the equipment historical normal operation data. When the new main frequency offset does not exceed the preset offset threshold value, the original operation parameter configuration of the equipment is maintained; when the new main frequency offset exceeds the threshold value, the system compares and matches the current new operation data (including vibration frequency, amplitude, load current) with the historical cases stored in the abnormal state mapping table, and optimizes the content of the mapping table by using a reinforcement learning algorithm based on Q learning. The algorithm takes the equipment vibration state as the environment state, the parameter adjustment action as the agent behavior, and the vibration amplitude reduction degree and the parameter adjustment amplitude as the compound reward function, and optimizes the corresponding relationship between the abnormal state and the parameter adjustment scheme by iteratively updating the Q value table. The new parameter configuration generated after each optimization is immediately applied to the equipment, and the adjustment effect is recorded, forming a closed-loop learning mechanism. The finally output optimized operation parameter configuration not only considers the current vibration characteristics, but also integrates historical adjustment experience, so that the parameter adjustment strategy has the ability of continuous improvement. This step introduces the reinforcement learning method, so that the system can adaptively improve the abnormal handling strategy and gradually improve the accuracy and effectiveness of parameter adjustment.

[0044] Referring to Figure 2 The second embodiment of the present application provides an intelligent control system for aluminum alloy profile production, comprising: A data acquisition module is configured to acquire operation data, wherein the operation data comprises vibration frequency and amplitude data. A feature analysis module is configured to extract initial vibration signal features based on the vibration frequency and the amplitude data by using a Fourier transform algorithm, wherein the initial vibration signal features comprise periodic characteristics, abnormal peak value characteristics, and a main frequency offset. An abnormality analysis module is configured to analyze and determine the vibration abnormality degree of a key component when the main frequency offset exceeds a preset offset threshold value, to obtain an abnormal offset evaluation result. A health state classification module is configured to input the abnormal offset evaluation result and the initial vibration signal features into a pre-established health state classifier to obtain a health state label. An operation parameter generation module is configured to generate an initial operation parameter configuration through a pre-established abnormal state mapping table when the health state label shows abnormality. A control operation module is configured to update the equipment operation state through a real-time control system according to the initial operation parameter configuration to obtain new operation data. The running parameter optimization module is configured to extract a new main frequency offset based on the new running data, and update the abnormal state mapping table to generate an optimized running parameter configuration when the new main frequency offset exceeds a preset offset threshold.

[0045] It should be noted that the intelligent control device for aluminum alloy profile production provided by the embodiments of the present application is used to execute all process steps of the intelligent control method for aluminum alloy profile production provided by the above embodiments, and the working principles and beneficial effects of the two are one-to-one correspondence, thus not being described in detail.

[0046] The embodiments of the present application further provide an electronic device. The electronic device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, for example, an intelligent control program for aluminum alloy profile production. The processor implements the steps in the above various embodiments of the intelligent control method for aluminum alloy profile production when executing the computer program, for example Figure 1 The processor implements the functions of the modules / units in the above various device embodiments when executing the computer program, for example, an intelligent control module for aluminum alloy profile production.

[0047] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the electronic device.

[0048] The electronic device can be a desktop computer, a notebook, a palm computer, and a smart tablet, etc. The electronic device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the above components are only examples of the electronic device, and do not constitute a limitation on the electronic device, and can include more or less components than the above, or combine certain components, or different components, for example, the electronic device can also include an input / output device, a network access device, a bus, etc.

[0049] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The processor is a control center of the electronic device, and connects various parts of the electronic device through various interfaces and lines.

[0050] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), etc.; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory device.

[0051] The modules / units integrated in the electronic device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. that can carry the computer program code. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0052] It should be noted that the above-described device embodiments are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. In addition, the connection relationship between the modules in the device embodiment provided by the present application indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0053] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above-described specific embodiments are only for the specific embodiments of the present application and are not used to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An intelligent control method for aluminum alloy profile production, characterized in that: include: collecting operating data, wherein the operating data includes vibration frequency and amplitude data; Extracting initial vibration signal features based on the vibration frequency and the amplitude data using a Fourier transform algorithm, wherein the initial vibration signal features include periodic features, abnormal peak features, and a main frequency offset; When the main frequency offset exceeds a preset offset threshold, analyzing and determining the degree of vibration abnormality of key components to obtain an abnormal offset assessment result; Inputting the abnormal deviation assessment result and the initial vibration signal feature into a pre-established health state classifier to obtain a health state label; When the health status tag indicates an abnormality, an initial operating parameter configuration is generated through a pre-established abnormal status mapping table; According to the initial operating parameter configuration, the device operating status is updated through the real-time control system to obtain new operating data; A new main frequency offset is extracted according to the new operating data. When the new main frequency offset exceeds a preset offset threshold, the abnormal state mapping table is updated to generate an optimized operating parameter configuration.

2. The intelligent control method for aluminum alloy profile production according to claim 1, characterized in that: The extracting of initial vibration signal features by a Fourier transform algorithm based on the vibration frequency and the amplitude data includes: Performing spectrum decomposition using a Fourier transform algorithm based on the vibration frequency and the amplitude data to obtain a frequency domain feature set; Identify and extract the frequency with the maximum amplitude as the main frequency according to the frequency domain feature set, and calculate the offset between the main frequency and the preset base frequency to obtain the main frequency offset; According to the frequency domain feature set, a sliding window technique is used to perform time domain analysis on the frequency domain feature set to obtain a periodic feature; When there is an abnormal point in the periodic feature whose amplitude data exceeds a preset amplitude threshold, performing local feature decomposition on the abnormal point by discrete wavelet transform to determine an abnormal peak feature; When there is no abnormal point in the periodic feature where the amplitude data exceeds a preset amplitude threshold, there is no abnormal peak feature; The main frequency offset, the periodic characteristics and the abnormal peak characteristics are used as initial vibration signal characteristics.

3. The intelligent control method for aluminum alloy profile production according to claim 1, characterized in that: When the main frequency offset exceeds a preset offset threshold, analyzing and determining the abnormal vibration degree of key components to obtain an abnormal offset assessment result includes: When the main frequency offset exceeds a preset offset threshold, a cosine similarity algorithm is used to calculate a similarity score between the periodic feature and pre-stored historical normal data, and the score is used as the degree of vibration abnormality; Performing weighted fusion on the vibration abnormality degree and the main frequency offset to generate an abnormality score; The mean filtering algorithm is used to smooth the anomaly score and output the anomaly offset assessment result.

4. The intelligent control method for aluminum alloy profile production according to claim 1, characterized in that: The process of constructing the health status classifier includes: Obtain historical vibration signal characteristics and corresponding historical health status; The support vector machine algorithm is used to construct the initial health status classifier, the classification boundary is initialized by the RBF kernel function, and the penalty coefficient is set to determine the tolerance error threshold; Based on the historical vibration signal characteristics, a kernel density estimation method is used to calculate the curvature adjustment parameter of the classification boundary, and the curvature of the classification boundary is dynamically adjusted through the RBF kernel function; According to the historical health status, the penalty coefficient is optimized by a grid search method to determine an error tolerance threshold; Perform cross-validation on the adjusted classification boundaries and calculate the classification accuracy and error distribution set after boundary adjustment; When the error distribution set exceeds the error tolerance threshold, the curvature parameter of the classification boundary is re-optimized using the gradient descent method, and the classification boundary is iteratively updated until the error distribution set does not exceed the error tolerance threshold; The optimized classification boundary and health status are matched in multi-dimensional space through RBF kernel function mapping to determine the hyperplane equation corresponding to each health status; Combined with the time series characteristics of the historical health status, the sliding window method is used to extract the dynamic trend parameters of the health status change, and the dynamic trend parameters are associated with the hyperplane equation to construct the final health status classifier.

5. The intelligent control method for aluminum alloy profile production according to claim 1, characterized in that: When the health status tag indicates an abnormality, generating an initial operating parameter configuration through a pre-established abnormal status mapping table includes: When the health status tag indicates abnormality, the load current of the device is obtained, and a correlation coefficient between the load current and the main frequency offset is calculated using a Pearson correlation coefficient algorithm to obtain a load correlation coefficient; When the load correlation coefficient exceeds a preset load correlation threshold, querying a pre-established abnormal state mapping table to generate an initial operating parameter configuration; When the load correlation coefficient does not exceed the preset load correlation threshold, the original operating parameter configuration is maintained to obtain the initial operating parameter configuration.

6. The intelligent control method for aluminum alloy profile production according to claim 1, characterized in that: The method of updating the device operating status by a real-time control system according to the initial operating parameter configuration to obtain new operating data includes: According to the initial operating parameter configuration, the corresponding adjustable parameter range is obtained from a pre-established operating parameter database to determine an adapted adjustment amplitude range; According to the initial operating parameter configuration and the adjustment amplitude range, real-time control instructions are constructed through the industrial transmission protocol to update the equipment operating status and obtain new operating data.

7. The intelligent control method for aluminum alloy profile production according to claim 1, characterized in that: The extracting a new main frequency offset according to the new operating data, and when the new main frequency offset exceeds a preset offset threshold, updating the abnormal state mapping table to generate an optimized operating parameter configuration includes: Performing spectrum analysis on the new operating data using a Fourier transform algorithm to obtain a new main frequency offset; When the new main frequency offset exceeds a preset offset threshold, the abnormal state mapping table is optimized using a reinforcement learning algorithm in combination with new operating data to generate an optimized operating parameter configuration.

8. An intelligent control system for aluminum alloy profile production, characterized in that: include: A data acquisition module, configured to acquire operating data, wherein the operating data includes vibration frequency and amplitude data; a feature analysis module, configured to extract initial vibration signal features using a Fourier transform algorithm based on the vibration frequency and the amplitude data, wherein the initial vibration signal features include periodic features, abnormal peak features, and a main frequency offset; an abnormality analysis module, configured to analyze and determine the degree of vibration abnormality of key components when the main frequency offset exceeds a preset offset threshold, and obtain an abnormal offset assessment result; a health status classification module, configured to input the abnormal deviation assessment result and the initial vibration signal feature into a pre-established health status classifier to obtain a health status label; An operating parameter generation module is used to generate an initial operating parameter configuration through a pre-established abnormal state mapping table when the health status label is displayed as abnormal; A control operation module is used to update the equipment operation status through a real-time control system according to the initial operation parameter configuration to obtain new operation data; The operating parameter optimization module is used to extract a new main frequency offset according to the new operating data, and when the new main frequency offset exceeds a preset offset threshold, update the abnormal state mapping table to generate an optimized operating parameter configuration.

9. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the intelligent control method for aluminum alloy profile production as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the intelligent control method for aluminum alloy profile production according to any one of claims 1 to 7.

Images