An abnormal vibration detection method, device, equipment and medium
By constructing a predictive model to detect abnormal vibrations using historical mill operating data, the problem of rapid detection of abnormal vibrations in cold rolling production was solved, enabling timely intervention and quality improvement.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 武汉钢铁有限公司
- Filing Date
- 2023-06-26
- Publication Date
- 2026-05-29
Smart Images
Figure CN116689514B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of rolling technology, and in particular to an abnormal vibration detection method, apparatus, equipment, and medium. Background Technology
[0002] During the cold rolling process, abnormal vibrations may sometimes occur in the rolling mill. These abnormal vibrations can cause vibration marks on the surface of the rolls and strip, accompanied by loud noise. If not addressed in time, they may lead to serious production accidents such as roll breakage or strip breakage.
[0003] In related technologies, the mill is typically slowed down after abnormal vibrations are detected manually. However, this is only a passive method of vibration suppression, and the optimal speed reduction window is often missed during actual production, leading to limited production capacity. Therefore, there is an urgent need for a technology that can quickly detect abnormal vibrations and intervene in a timely manner to reduce the negative impact on the mill and the production process. Summary of the Invention
[0004] This application provides an abnormal vibration detection method, device, equipment, and medium, which solves the technical problem of the inability to quickly detect abnormal vibrations in the prior art. It achieves the technical effect of quickly detecting abnormal vibrations, thereby enabling timely intervention and reducing the negative impact on the rolling mill and production process.
[0005] Firstly, this application provides a method for detecting abnormal vibration, the method comprising:
[0006] Acquire historical operating data of the target rolling mill, including vibration measurement data, rolled sheet shape data, rolled sheet production data, and rolling process data that correspond one-to-one in time sequence;
[0007] Training and testing sample sets are constructed based on historical operational data;
[0008] The target prediction model is obtained by training the pre-set prediction model based on the training sample set and the test sample set.
[0009] The real-time operating data of the mill under test is obtained and input into the target prediction model to determine whether there is abnormal vibration in the mill under test.
[0010] Furthermore, the rolling production data includes at least one of the following: rolling width, thickness, roll diameter, and roll roughness; the rolling process data includes at least one of the following: rolling speed, working tension, rolling force, emulsion parameters, and roll gap of the target mill during historical production processes.
[0011] Furthermore, training and testing sample sets are constructed based on historical operational data, including:
[0012] Frequency coordination is performed on historical operational data to obtain preprocessed historical data;
[0013] Training and testing sample sets are constructed based on preprocessed historical data.
[0014] Furthermore, frequency coordination is performed on historical operational data to obtain preprocessed historical data, including:
[0015] Based on the collection frequency of rolling process data in historical operation data, vibration measurement data and rolled piece shape data are downsampled, and rolled piece production data is augmented.
[0016] The rolling process data, downsampled vibration measurement data, rolled piece shape data, and expanded rolled piece production data were normalized to obtain preprocessed historical data.
[0017] Furthermore, before performing frequency coordination on historical operational data, the method also includes:
[0018] Vibration energy data is determined based on vibration measurement data.
[0019] Furthermore, historical operating data refers to the operating data generated by the target rolling mill during a continuous rolling period with a preset duration.
[0020] Furthermore, real-time operational data is input into the target prediction model to determine whether abnormal vibrations exist in the rolling mill under test, including:
[0021] Input the real-time running data into the target prediction model to obtain the vibration energy prediction value;
[0022] When the predicted vibration energy value exceeds the preset vibration energy value, it is determined that there is abnormal vibration in the rolling mill under test.
[0023] Secondly, this application provides an abnormal vibration detection device, the device comprising:
[0024] The acquisition module is used to acquire historical operating data of the target rolling mill. The historical operating data includes vibration measurement data, rolled piece shape data, rolled piece production data and rolling process data that correspond one-to-one in time sequence.
[0025] The building module is used to construct training and testing sample sets based on historical running data;
[0026] The training module is used to train the preset prediction model based on the training sample set and the test sample set to obtain the target prediction model.
[0027] The detection module is used to acquire real-time operating data of the mill under test and input the real-time operating data into the target prediction model to determine whether there is abnormal vibration in the mill under test.
[0028] Thirdly, this application provides an electronic device, comprising:
[0029] processor;
[0030] Memory used to store processor-executable instructions;
[0031] The processor is configured to execute an abnormal vibration detection method as provided in the first aspect.
[0032] Fourthly, this application provides a non-transitory computer-readable storage medium that, when the instructions in the storage medium are executed by the processor of an electronic device, enables the electronic device to perform an abnormal vibration detection method as provided in the first aspect.
[0033] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:
[0034] This application embodiment constructs a target prediction model using historical operating data of the rolling mill. Based on this model, it detects the actual operating data of the rolling mill under test to determine whether abnormal vibrations exist. Therefore, this application embodiment, based on big data modeling and prediction, avoids the complex conditions of strong coupling, nonlinearity, and polymorphism between variables in traditional process mechanism theory modeling. By training and learning from production data, a data-driven non-mechanistic model is established, enabling dynamic monitoring and prediction of vibration marks. In other words, this application embodiment can identify the optimal speed reduction window before abnormal vibrations cause significant damage to the rolls or rolled pieces, thereby allowing for timely control of the rolling mill to reduce production speed, decrease the probability of abnormal vibrations, improve the rolling quality of the rolled pieces, and extend the service life of the rolling mill to some extent. Attached Figure Description
[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0036] Figure 1 A flowchart illustrating an abnormal vibration detection method provided in this application;
[0037] Figure 2 Applications provided for this application Figure 1 A schematic diagram of the detection system of the method shown;
[0038] Figure 3 This is a schematic diagram showing the sample features and sample labels contained in a given sample.
[0039] Figure 4 A region indication map for predicting abnormal vibrations as shown in the human-computer interaction interface;
[0040] Figure 5 This is an area indication map for predicting abnormal vibrations in another human-computer interaction interface;
[0041] Figure 6 This is a schematic diagram of the structure of an abnormal vibration detection device provided in this application;
[0042] Figure 7 This is a schematic diagram of the structure of an electronic device provided in this application. Detailed Implementation
[0043] This application provides an abnormal vibration detection method, which solves the technical problem in the prior art that abnormal vibrations cannot be detected quickly.
[0044] The technical solution of this application embodiment is to solve the above-mentioned technical problems, and the general idea is as follows:
[0045] An abnormal vibration detection method includes: acquiring historical operating data of a target rolling mill, including vibration measurement data, rolled sheet shape data, rolled sheet production data, and rolling process data that correspond one-to-one in time sequence; constructing a training sample set and a test sample set based on the historical operating data; training a preset prediction model based on the training sample set and the test sample set to obtain a target prediction model; acquiring real-time operating data of the rolling mill to be detected and inputting the real-time operating data into the target prediction model to determine whether the rolling mill to be detected has abnormal vibration.
[0046] This application embodiment constructs a target prediction model using historical operating data of the rolling mill. Based on this model, it detects the actual operating data of the rolling mill under test to determine whether abnormal vibrations exist. Therefore, this application embodiment, based on big data modeling and prediction, avoids the complex conditions of strong coupling, nonlinearity, and polymorphism between variables in traditional process mechanism theory modeling. By training and learning from production data, a data-driven non-mechanistic model is established, enabling dynamic monitoring and prediction of vibration marks. In other words, this application embodiment can identify the optimal speed reduction window before abnormal vibrations cause significant damage to the rolls or rolled pieces, thereby allowing for timely control of the rolling mill to reduce production speed, decrease the probability of abnormal vibrations, improve the rolling quality of the rolled pieces, and extend the service life of the rolling mill to some extent.
[0047] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.
[0048] First, it should be clarified that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.
[0049] During cold rolling production, abnormal vibrations may occur in the rolling mill. These abnormal vibrations cause oscillation marks on the rolls and workpiece surfaces, accompanied by significant noise. With the continuous increase in rolling mill production line speed, the probability of abnormal vibrations during rolling is increasing, leading to a significant increase in the incidence of oscillation mark defects. Oscillation marks have become a bottleneck restricting the production efficiency and product quality improvement of pickling and rolling mills.
[0050] Normally, after abnormal vibration occurs, speed reduction is required. Failure to reduce speed may lead to serious production accidents such as roll breakage or strip breakage. However, currently, speed reduction is usually implemented only after abnormal vibration is detected manually. This only passively suppresses vibration, and the optimal speed reduction window is often missed during actual production. This results in the inability to intervene in the impact of abnormal vibration on the rolls and rolled pieces in a timely manner, and also fails to guarantee rolling efficiency, greatly limiting production capacity.
[0051] In addition, due to the different processes and the complexity of on-site conditions, existing technologies can rely on traditional vibration mechanism models. However, the variables in the vibration mechanism model are subject to complex conditions such as strong coupling, nonlinearity, and polymorphism, making it difficult to truly reflect the actual situation of abnormal vibration in the rolling mill.
[0052] Therefore, there is an urgent need for a technology that can quickly detect abnormal vibrations and intervene in them in a timely manner to reduce the negative impact on the rolling mill and production process.
[0053] To address the aforementioned problems, this embodiment provides the following... Figure 1 The method shown includes steps S11-S14.
[0054] Step S11: Obtain historical operating data of the target rolling mill. The historical operating data includes vibration measurement data, rolled sheet shape data, rolled sheet production data and rolling process data that correspond one-to-one in time sequence.
[0055] Step S12: Construct training sample sets and test sample sets based on historical running data;
[0056] Step S13: Train the preset prediction model based on the training sample set and the test sample set to obtain the target prediction model;
[0057] Step S14: Obtain the real-time operating data of the mill to be tested, and input the real-time operating data into the target prediction model to determine whether there is abnormal vibration in the mill to be tested.
[0058] Regarding step S11, the historical operating data of the target rolling mill is obtained. The historical operating data includes vibration measurement data, rolled piece shape data, rolled piece production data and rolling process data that correspond one-to-one in time sequence.
[0059] Historical operating data refers to the valid operating data generated by the target rolling mill during a continuous rolling period of a preset duration in its historical production process. The longer the preset duration, the more operating data is obtained, resulting in a larger number of samples in the training and testing sample sets, and ultimately, a higher accuracy of the trained target prediction model. Therefore, the preset duration can be set according to actual conditions. When higher accuracy is required for the target prediction model, a longer preset duration is set; when lower accuracy is required, a shorter preset duration can be selected.
[0060] like Figure 2 The diagram shown is a schematic of the structure of a prediction system provided in this embodiment. The diagram includes mill vibration signals (i.e., vibration measurement data) and sheet shape data (i.e., rolled sheet shape data), as well as a secondary server in the factory (used to acquire rolled sheet production data) and a field PLC control system (used to acquire rolling process data).
[0061] Vibration measurement data includes actual vibration data of the target rolling mill during historical production processes, which can be acquired using vibration acceleration sensors installed on the rolling mill stand. The vibration measurement data acquired by the vibration acceleration sensor is a vibration time-domain signal. In this embodiment, the vibration energy value can be determined first through the vibration time-domain signal. In the subsequent step S12, the vibration measurement data in the form of vibration energy is used to construct the sample space (the sample space includes a training sample set and a test sample set).
[0062] The shape data of the rolled piece can be obtained by a shape meter installed at the mill exit.
[0063] The rolling mill production data includes at least one of the following: rolling mill width, thickness, roll diameter, and roll roughness. Specifically, it can be obtained through communication with the secondary machine server corresponding to the rolling mill.
[0064] Rolling process data includes at least one of the following in the historical production process of the target mill: rolling speed, working tension, rolling force, emulsion parameters, and roll gap. Specifically, it can be obtained through communication with the PLC (Programmable Logic Controller) on site.
[0065] The vibration measurement data, rolled piece shape data, rolled piece production data, and rolling process data that correspond one-to-one in time sequence refer to the vibration measurement data, rolled piece shape data, rolled piece production data, and rolling process data generated at the same moment or within the same time period that correspond to each other.
[0066] Regarding step S12, a training sample set and a test sample set are constructed based on historical running data.
[0067] Vibration measurement data, roll shape data, roll production data, and rolling process data from historical operation data are collected by different physical systems. The data collection frequencies vary, and some data have large dimensions. Therefore, frequency-coordinated preprocessing is required. Specifically, this includes: frequency coordination of historical operation data to obtain preprocessed historical data; and constructing training and testing sample sets based on the preprocessed historical data.
[0068] Frequency coordination is performed on historical operational data to obtain preprocessed historical data, including:
[0069] Based on the collection frequency of rolling process data in historical operation data, vibration measurement data and rolled piece shape data are downsampled, and rolled piece production data is augmented.
[0070] The rolling process data, downsampled vibration measurement data, rolled piece shape data, and expanded rolled piece production data were normalized to obtain preprocessed historical data.
[0071] The sampling frequency of rolling process data is the lowest; therefore, it is necessary to downsample the vibration measurement data and rolled piece shape data based on the lowest sampling frequency of rolling process data. In addition, it is necessary to expand the rolled piece production data. As shown in Table 1, these are the sampling frequencies of various data in a certain historical operating data set. The sampling frequency of rolling process data is the lowest; therefore, the vibration measurement data and rolled piece shape data are downsampled based on the sampling frequency of the rolling process data, i.e., the sampling frequency of vibration measurement data and rolled piece shape data is reduced (i.e., some data is discarded). Furthermore, the rolled piece production data is expanded.
[0072] Table 1
[0073] Data types Sampling interval (s) Sampling frequency (Hz) Vibration measurement data 0.2 5 Rolling process data 5-7 0.14-0.2 Rolled product production data Each roll of steel Each roll of steel Rolled sheet shape data 0.072 15
[0074] In practice, it can be carried out in the following manner:
[0075] (1) Data dimensionality reduction. Strip shape data is collected by a strip shape meter at the exit of the rolling mill stand. The strip shape meter records the stress values of each strip along the width direction at regular intervals. Since the strip shape meter has a large number of sensors, generally dozens to hundreds, the corresponding stress value parameters are numerous. To avoid the "dimensionality explosion" defect, a dimensionality reduction method is adopted.
[0076] For example, in a certain system, for each frame of plate shape data zone1 to zonen, a fifth-order Legendre orthogonal polynomial is selected for fitting, and the fitting polynomial coefficients for each frame are obtained: (C0, C1, C2, C3, C4), which can be used as the characteristic values of the plate shape.
[0077] (2) Unify the data acquisition frequency. As shown in Table 1, the historical data collected by the factory comes from different sources and needs to be preprocessed with frequency coordination before it can be transformed into sample data for model learning and training. Based on the acquisition frequency of rolling process parameters, the vibration energy and plate shape data are downsampled, and the production data is expanded to achieve the unification of the sampling frequency of each sample.
[0078] (3) Normalization. The varying scales of different data units can affect the iteration speed and the weights between parameters during training. Therefore, data standardization is necessary. This technical solution employs normalization, normalizing each feature to the [0,1] interval. The normalization formula is as follows:
[0079]
[0080] Where x new Here, x represents the normalized sample feature values, and x represents the sample feature values before normalization. min x is the minimum value of the sample data for this feature. max This represents the maximum value of the sample data for this feature.
[0081] After processing the data as described above, preprocessed historical data is obtained. Then, training and testing sample sets are constructed based on this preprocessed historical data. Specifically, a portion of the preprocessed historical data is used as training samples, and the other portion is used as testing samples, forming the training and testing sample sets respectively. The ratio of the number of samples in the training and testing sample sets can be set according to the actual situation. Typically, the number of samples in the training sample set is greater than the number of samples in the testing sample set; therefore, a ratio of 3:1 is suitable.
[0082] The training sample set and the test sample set each include sample labels and sample features. The sample labels are vibration energy data (i.e. vibration measurement data), and the sample features are rolled piece shape data, rolled piece production data and rolling process data. The sample labels and sample features correspond to each other.
[0083] For example, such as Figure 3 As shown, each sample contains 14 feature values, and the sample label is the vibration energy value. The feature values and label values collected in each sampling period constitute a sample vector, and all valid sample vectors collected during the continuous rolling historical time constitute the sample space.
[0084] Regarding step S13, the preset prediction model is trained based on the training sample set and the test sample set to obtain the target prediction model.
[0085] A training sample set is input into a pre-defined prediction model to train it, resulting in a trained prediction model to be tested. Then, a test sample set is input into this model to determine if its accuracy meets a pre-defined requirement. If the accuracy does not meet the requirement, the model is retrained using the training sample set, or a new training sample set can be used to train it, resulting in a new prediction model. The accuracy of this new model is then assessed to determine if it meets the pre-defined requirement. If it does, this new model is used as the target prediction model.
[0086] For example, 75% of the data in the sample space is used as the training sample set, and 25% is used as the training sample set for modeling and testing. The prediction algorithm uses a deep learning neural network algorithm to calculate the target prediction model for abnormal vibration of the strip steel.
[0087] In actual production, as production continues, the on-site operating conditions and environmental parameters change, causing the accuracy of the old model (i.e., the previously obtained target prediction model) to continuously decrease in its prediction accuracy for input data under new operating conditions. Therefore, the model can be retrained according to a preset cycle, i.e., the target prediction model can be updated, or the model can be retrained based on the degree of change in on-site operating conditions and environmental parameters. Retraining the model can rely on recently generated historical operating data, i.e., steps S11-S13 can be re-executed using recently generated historical operating data to obtain an updated target prediction model. Here, "recently generated historical operating data" refers to data generated within a preset operating period after the last acquisition of the target prediction model.
[0088] Regarding step S14, real-time operating data of the mill to be tested is obtained, and the real-time operating data is input into the target prediction model to determine whether there is abnormal vibration in the mill to be tested.
[0089] The mill under test and the target mill can be the same mill. In this case, the accuracy is highest when the target prediction model is trained using the mill's own operating data, and then the abnormal vibration of the mill itself is detected. Of course, the mill under test can also be the same type of mill as the target mill, which also results in a high accuracy in detecting abnormal vibration.
[0090] In addition, if high accuracy is not required, other types of rolling mills besides the target rolling mill can be used as the rolling mill to detect whether there is abnormal vibration.
[0091] Real-time operational data includes vibration measurement data, rolled piece shape data, rolled piece production data, and rolling process data that correspond one-to-one in time sequence.
[0092] Real-time operating data is input into the target prediction model, and the target prediction model outputs the corresponding vibration energy prediction value; when the vibration energy prediction value exceeds the vibration energy preset value, it is determined that there is abnormal vibration in the mill to be tested.
[0093] For example, the preset vibration energy value is 125% of the standard vibration energy corresponding to the normal operation of the rolling mill. If the vibration energy prediction value output by the target prediction model exceeds 125%, it is considered that there is abnormal vibration in the rolling mill under test, and a warning signal can be issued. The warning signal can directly trigger the corresponding control equipment to control the rolling mill to reduce speed, or it can only issue a prompt message, which can promptly notify the operator to control the rolling mill to reduce speed. This can greatly reduce the probability of abnormal vibration causing vibration marks on the rolls and rolled products, improve product quality, and extend the production line life.
[0094] Real-time operating data can be the operating data of a portion of a coil of strip steel, for example, it can be the data of one-tenth of the length of a coil of strip steel. This means that the target prediction model can determine whether there is abnormal vibration in each part of the strip steel, which can further reduce the probability of abnormal vibration causing vibration marks on the rolls and rolled pieces, improve product quality, and extend the production line life.
[0095] like Figure 4 and Figure 5 The diagram shows the human-machine interface provided in this embodiment for actual use. It allows for the identification of vibration signal areas, process parameter areas, and abnormal vibration areas, facilitating operators to quickly pinpoint abnormal vibration regions and intervene in the damage to the rolls or rolled pieces by reducing speed within a short time. The light gray area represents the rolled strip area, the medium gray area represents the abnormal vibration area, and the dark gray area represents the unrolled strip area. Additionally, in... Figure 5 The system provides a model update button, which allows you to retrain the target prediction model to match the actual operating conditions of the rolling mill and ensure the accuracy of the target prediction model.
[0096] In summary, this embodiment constructs a target prediction model using historical operating data of the rolling mill. Based on this model, it detects the actual operating data of the rolling mill under test to determine whether abnormal vibrations exist. It is evident that this embodiment, based on big data modeling and prediction, avoids the complex conditions of strong coupling, nonlinearity, and polymorphism between variables in traditional process mechanism theory modeling. By training and learning from production data, a data-driven non-mechanistic model is established, enabling dynamic monitoring and prediction of vibration marks. In other words, this embodiment can identify the optimal speed reduction window before abnormal vibrations cause significant damage to the rolls or rolled pieces, thereby allowing for timely control of the rolling mill to reduce production speed, decrease the probability of abnormal vibrations, improve the rolling quality of the rolled pieces, and extend the service life of the rolling mill to some extent.
[0097] Based on the same inventive concept, this embodiment provides as follows: Figure 6 An abnormal vibration detection device is shown, the device comprising:
[0098] The acquisition module 61 is used to acquire the historical operating data of the target rolling mill. The historical operating data includes vibration measurement data, rolled piece shape data, rolled piece production data and rolling process data that correspond one-to-one in time sequence.
[0099] Module 62 is used to build training and testing sample sets based on historical running data;
[0100] Training module 63 is used to train a preset prediction model based on the training sample set and the test sample set to obtain the target prediction model;
[0101] The detection module 64 is used to acquire real-time operating data of the mill to be inspected and input the real-time operating data into the target prediction model to determine whether there is abnormal vibration in the mill to be inspected.
[0102] Furthermore, the rolling production data includes at least one of the following: rolling width, thickness, roll diameter, and roll roughness; the rolling process data includes at least one of the following: rolling speed, working tension, rolling force, emulsion parameters, and roll gap of the target mill during historical production processes.
[0103] Furthermore, module 62 includes:
[0104] The frequency coordination submodule is used to perform frequency coordination on historical operating data to obtain preprocessed historical data;
[0105] The construction submodule is used to build training and testing sample sets based on preprocessed historical data.
[0106] Furthermore, the frequency coordination submodule includes:
[0107] The data preprocessing submodule is used to downsample the vibration measurement data and the shape data of the rolled piece based on the acquisition frequency of the rolling process data in the historical operation data, and to expand the data of the rolled piece production data.
[0108] The normalization submodule is used to normalize rolling process data, downsampled vibration measurement data, rolled piece shape data, and expanded rolled piece production data to obtain preprocessed historical data.
[0109] Furthermore, the device also includes a conversion module for determining vibration energy data based on vibration measurement data before frequency coordination of historical operating data.
[0110] Furthermore, historical operating data refers to the operating data generated by the target rolling mill during a continuous rolling period with a preset duration.
[0111] Furthermore, the detection module 64 is specifically used for:
[0112] Input the real-time running data into the target prediction model to obtain the vibration energy prediction value;
[0113] When the predicted vibration energy value exceeds the preset vibration energy value, it is determined that there is abnormal vibration in the rolling mill under test.
[0114] Based on the same inventive concept, this embodiment provides as follows: Figure 7 An electronic device shown includes:
[0115] Processor 71;
[0116] Memory 72 is used to store executable instructions of processor 71;
[0117] The processor 71 is configured to execute an abnormal vibration detection method as described above.
[0118] Based on the same inventive concept, this embodiment provides a non-transitory computer-readable storage medium, which, when the instructions in the storage medium are executed by the processor 71 of the electronic device, enables the electronic device to perform an abnormal vibration detection method as described above.
[0119] Since the electronic device described in this embodiment is an electronic device used to implement the information processing method in the embodiments of this application, those skilled in the art can understand the specific implementation methods and various variations of the electronic device in this embodiment based on the information processing method described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here. Any electronic device used by those skilled in the art to implement the information processing method in the embodiments of this application falls within the scope of protection of this application.
[0120] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0121] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0122] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0123] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0124] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.
[0125] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.
Claims
1. A method for detecting abnormal vibration, characterized in that, The method includes: Historical operating data of the target rolling mill is acquired. This historical operating data includes, in a time-series correspondence, vibration measurement data, workpiece shape data, workpiece production data, and rolling process data. The vibration measurement data includes actual vibration data of the target rolling mill during historical production processes. The workpiece shape data is collected by a shape gauge at the mill stand exit, which records the stress values of each shape along the width direction of the workpiece. The workpiece production data includes at least one of the following: workpiece width, thickness, roll diameter, and roll roughness. The rolling process data includes at least one of the following: rolling speed, working tension, rolling force, emulsion parameters, and roll gap during the historical production processes of the target rolling mill. A training sample set and a test sample set are constructed based on the historical operating data; The preset prediction model is trained based on the training sample set and the test sample set to obtain the target prediction model; The real-time operating data of the mill to be tested is acquired and input into the target prediction model to determine whether the mill to be tested has abnormal vibrations. The construction of the training sample set and test sample set based on the historical operating data includes: Frequency coordination is performed on the historical operation data to obtain preprocessed historical data; The training sample set and the test sample set are constructed based on the preprocessed historical data; The process of frequency coordination of the historical operational data to obtain preprocessed historical data includes: Based on the acquisition frequency of the rolling process data in the historical operating data, the vibration measurement data and the rolled piece shape data are downsampled, and the rolled piece production data is augmented. The rolling process data, downsampled vibration measurement data, rolled piece shape data, and expanded rolled piece production data are normalized to obtain the preprocessed historical data.
2. The method as described in claim 1, characterized in that, Before performing frequency coordination on the historical operational data, the method further includes: The vibration energy data is determined based on the vibration measurement data.
3. The method as described in claim 1, characterized in that, The historical operating data refers to the operating data generated by the target rolling mill during a continuous rolling period with a preset duration.
4. The method as described in claim 1, characterized in that, The step of inputting the real-time operating data into the target prediction model to determine whether the mill under test has abnormal vibration includes: The real-time operating data is input into the target prediction model to obtain the vibration energy prediction value; When the predicted vibration energy value exceeds the preset vibration energy value, it is determined that the mill under test has abnormal vibration.
5. An abnormal vibration detection device, characterized in that, The device for detecting abnormal vibrations according to any one of claims 1 to 4 comprises: The acquisition module is used to acquire historical operating data of the target rolling mill, including vibration measurement data, rolled piece shape data, rolled piece production data and rolling process data that correspond one-to-one in time sequence; The construction module is used to construct training sample sets and test sample sets based on the historical running data; The training module is used to train a preset prediction model based on the training sample set and the test sample set to obtain a target prediction model. The detection module is used to acquire real-time operating data of the mill to be inspected and input the real-time operating data into the target prediction model to determine whether the mill to be inspected has abnormal vibration; The building blocks include: The frequency coordination submodule is used to perform frequency coordination on historical operating data to obtain preprocessed historical data; The construction submodule is used to build training and testing sample sets based on preprocessed historical data. The frequency coordination submodule includes: The data preprocessing submodule is used to downsample the vibration measurement data and the shape data of the rolled piece based on the acquisition frequency of the rolling process data in the historical operation data, and to expand the data of the rolled piece production data. The normalization submodule is used to normalize rolling process data, downsampled vibration measurement data, rolled piece shape data, and expanded rolled piece production data to obtain preprocessed historical data.
6. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute an abnormal vibration detection method as described in any one of claims 1 to 4.
7. A non-transitory computer-readable storage medium, wherein instructions in the storage medium, when executed by a processor of an electronic device, enable the electronic device to perform an abnormal vibration detection method as described in any one of claims 1 to 4.