技术领域Technical field
本发明涉及数字孪生系统技术领域，尤其涉及一种热轧数字孪生带钢横截面形状实时预测方法及装置。The invention relates to the technical field of digital twin systems, and in particular to a method and device for real-time prediction of the cross-sectional shape of hot-rolled digital twin strips.
背景技术Background technique
带钢横断面形状是决定热轧带钢产品质量的重要指标之一。现场利用凸度仪对带钢横断面形状进行在线监测，并实现对断面形状的实时反馈控制，但是轧机辊缝前和在反馈控制不能投入的情况下仍无法保障带钢的带钢横断面质量。尤其对于热连轧，在带钢头部进入轧机之前，轧机的各板形调节机构都应有正确的预设定值，以保证闭环反馈控制模型投入工作前所轧带钢的板形，并作为闭环反馈控制的起点。预设定控制的精度不但关系到带钢的成品率，同时预设定值也是反馈控制的初始值，直接影响到板形反馈控制模块调整板形达到目标值的收敛速度和精度。The cross-sectional shape of the strip is one of the important indicators that determines the quality of the hot-rolled strip product. The crown meter is used to monitor the cross-section shape of the strip online on-site and realize real-time feedback control of the cross-section shape. However, the strip cross-section quality of the strip cannot be guaranteed before the roll gap of the rolling mill and when the feedback control cannot be put in. . Especially for hot continuous rolling, before the strip head enters the rolling mill, each plate shape adjustment mechanism of the rolling mill should have correct preset values to ensure the shape of the rolled strip before the closed-loop feedback control model is put into operation, and as the starting point for closed-loop feedback control. The accuracy of the preset control is not only related to the yield of the strip, but the preset value is also the initial value of the feedback control, which directly affects the convergence speed and accuracy of the flatness feedback control module to adjust the flatness to the target value.
发明内容Contents of the invention
本发明针对如何避免工艺参数设定不合理导致的热轧产品缺陷和降级现象的问题，提出了本发明。The present invention is directed at the problem of how to avoid defects and degradation of hot-rolled products caused by unreasonable setting of process parameters.
为解决上述技术问题，本发明提供如下技术方案：In order to solve the above technical problems, the present invention provides the following technical solutions:
一方面，本发明提供了一种热轧数字孪生带钢横截面形状实时预测方法，该方法由电子设备实现，该方法包括：On the one hand, the present invention provides a real-time prediction method for the cross-sectional shape of hot-rolled digital twin strip. The method is implemented by electronic equipment. The method includes:
S1、构建热连轧数字孪生产线，获取热连轧数字孪生产线的设定参数。S1. Construct a hot rolling digital twin production line and obtain the setting parameters of the hot rolling digital twin production line.
S2、采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，建立带钢横截面形状预测模型DMD-SINDy。S2. Use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm and establish the strip cross-sectional shape prediction model DMD-SINDy.
S3、根据设定参数以及带钢横截面形状预测模型，得到热连轧数字孪生产线的带钢横截面形状预测结果。S3. According to the set parameters and the strip cross-sectional shape prediction model, obtain the strip cross-sectional shape prediction results of the hot continuous rolling digital twin production line.
可选地，S2中的采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，建立带钢横截面形状预测模型DMD-SINDy，包括：Optionally, the dynamic mode decomposition DMD algorithm is used in S2 to optimize the nonlinear system dynamics sparse identification algorithm SINDy model and establish the strip cross-sectional shape prediction model DMD-SINDy, including:
S21、收集热连轧生产过程的历史数据，对历史数据进行预处理。S21. Collect historical data of the hot continuous rolling production process and preprocess the historical data.
其中，历史数据包括实测轧制力、弯辊力、两侧轧制力弯辊力差以及凸度实测值。Among them, the historical data includes the measured rolling force, roll bending force, rolling force and roll bending force difference on both sides, and measured crown value.
S22、建立非线性系统动力学稀疏识别算法SINDy模型。S22. Establish the SINDy model of the nonlinear system dynamics sparse identification algorithm.
S23、采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，得到带钢横截面形状预测模型DMD-SINDy。S23. Use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm, and obtain the strip cross-sectional shape prediction model DMD-SINDy.
可选地，S21中的对历史数据进行预处理，包括：Optionally, preprocessing historical data in S21 includes:
对历史数据进行线性变换，如下式（1）所示：Perform linear transformation on historical data, as shown in the following equation (1):
（1） (1)
其中，表示线性变换后的数据，/>表示线性变换前的数据，/>表示历史数据数量。in, Represents the data after linear transformation,/> Represents data before linear transformation,/> Indicates the amount of historical data.
可选地，S22中的建立非线性系统动力学稀疏识别算法SINDy模型，包括：Optionally, the nonlinear system dynamics sparse identification algorithm SINDy model established in S22 includes:
S221、根据预处理后的历史数据，构建时空矩阵和稀疏函数基库/>。S221. Construct a space-time matrix based on the preprocessed historical data. and sparse function base library/> .
S222、对时空矩阵和稀疏函数基库/>，经过稀疏回归，生成稀疏模型。S222. For space-time matrix and sparse function base library/> , through sparse regression, a sparse model is generated.
S223、根据序贯最小二乘回归法，确定稀疏解，根据稀疏模型以及稀疏解/>，建立非线性系统动力学稀疏识别算法SINDy模型。S223. Determine the sparse solution according to the sequential least squares regression method , according to the sparse model and sparse solution/> , establish the SINDy model of the nonlinear system dynamics sparse identification algorithm.
可选地，S221中的构建时空矩阵，包括：Optionally, construct the space-time matrix in S221 ,include:
对预处理后的历史数据，按照空间序列进行时空转换重采样，建立时空矩阵，如下（2）所示：For the preprocessed historical data, perform spatiotemporal conversion and resampling according to the spatial sequence to establish a spatiotemporal matrix. , as shown in (2) below:
（2） (2)
其中，表示空间序列，/>表示预处理后的历史数据，/>表示矩阵转置。in, Represents a spatial sequence,/> Represents preprocessed historical data,/> Represents matrix transpose.
可选地，S222中的稀疏模型，如下（3）所示：Optionally, the sparse model in S222 , as shown in (3) below:
（3） (3)
其中，表示稀疏函数基库，/>表示稀疏解。in, Represents a sparse function base library, /> represents a sparse solution.
可选地，S223中的非线性系统动力学稀疏识别算法SINDy模型，如下（4）所示：Optionally, nonlinear system dynamics sparse identification algorithm SINDy model in S223 , as shown in (4) below:
（4） (4)
其中，表示定义系统运动方程的动态约束，/>表示稀疏解，/>表示矩阵转置， 表示/>元素符号函数的向量。in, Represents dynamic constraints that define the system's equations of motion,/> Represents a sparse solution,/> represents the matrix transpose, Express/> Vector of element-wise symbolic functions.
可选地，S23中的采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，包括：Optionally, the dynamic mode decomposition DMD algorithm in S23 is used to optimize the nonlinear system dynamics sparse identification algorithm SINDy model, including:
S231、设定由预处理后的历史数据所组成的矩阵，矩阵中的元素包括随时间演变的第一时序数据向量以及与第一时序数据向量/>存在线性关系的第二时序数据向量/>，且/>，其中，/>为状态矩阵。S231. Set a matrix composed of preprocessed historical data. The elements in the matrix include the first time series data vector that evolves over time. and with the first timing data vector/> The second time series data vector with a linear relationship/> , and/> , where,/> is the state matrix.
S232、对第一时序数据向量进行简化的奇异值分解SVD，得到状态矩阵/>的模态。S232. For the first time series data vector Perform simplified singular value decomposition SVD to obtain the state matrix/> modal.
S233、根据简化的奇异值分解SVD以及状态矩阵的模态，得到任意时间点的历史数据。S233. Based on simplified singular value decomposition SVD and state matrix mode to obtain historical data at any point in time.
可选地，S233中的任意时间点的历史数据，如下（5）所示：Optionally, historical data at any point in time in S233 , as shown in (5) below:
（5） (5)
其中，表示模态的总数，/>表示本征向量，/>表示时间点t下的稀疏旋钮，/>表示各模式的模态振幅。in, Indicates the total number of modalities,/> represents the eigenvector,/> Represents the sparse knob at time point t, /> Represents the modal amplitude of each mode.
另一方面，本发明提供了一种热轧数字孪生带钢横截面形状实时预测装置，该装置应用于实现热轧数字孪生带钢横截面形状实时预测方法，该装置包括：On the other hand, the present invention provides a real-time prediction device for the cross-sectional shape of a hot-rolled digital twin strip. The device is used to implement a real-time prediction method for the cross-sectional shape of a hot-rolled digital twin strip. The device includes:
获取模块，用于构建热连轧数字孪生产线，获取热连轧数字孪生产线的设定参数。The acquisition module is used to construct the hot rolling digital twin production line and obtain the setting parameters of the hot rolling digital twin production line.
输入模块，用于采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，建立带钢横截面形状预测模型DMD-SINDy。The input module is used to use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm and establish the strip cross-sectional shape prediction model DMD-SINDy.
输出模块，用于根据设定参数以及带钢横截面形状预测模型，得到热连轧数字孪生产线的带钢横截面形状预测结果。The output module is used to obtain the strip cross-sectional shape prediction results of the hot continuous rolling digital twin production line based on the set parameters and the strip cross-sectional shape prediction model.
可选地，输入模块，进一步用于：Optionally, input modules are further used to:
S21、收集热连轧生产过程的历史数据，对历史数据进行预处理。S21. Collect historical data of the hot continuous rolling production process and preprocess the historical data.
其中，历史数据包括实测轧制力、弯辊力、两侧轧制力弯辊力差以及凸度实测值。Among them, the historical data includes the measured rolling force, roll bending force, rolling force and roll bending force difference on both sides, and measured crown value.
S22、建立非线性系统动力学稀疏识别算法SINDy模型。S22. Establish the SINDy model of the nonlinear system dynamics sparse identification algorithm.
S23、采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，得到带钢横截面形状预测模型DMD-SINDy。S23. Use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm, and obtain the strip cross-sectional shape prediction model DMD-SINDy.
可选地，输入模块，进一步用于：Optionally, input modules are further used to:
对历史数据进行线性变换，如下式（1）所示：Perform linear transformation on historical data, as shown in the following equation (1):
（1） (1)
其中，表示线性变换后的数据，/>表示线性变换前的数据，/>表示历史数据数量。in, Represents the data after linear transformation,/> Represents data before linear transformation,/> Indicates the amount of historical data.
可选地，输入模块，进一步用于：Optionally, input modules are further used to:
S221、根据预处理后的历史数据，构建时空矩阵和稀疏函数基库/>。S221. Construct a space-time matrix based on the preprocessed historical data. and sparse function base library/> .
S222、对时空矩阵和稀疏函数基库/>，经过稀疏回归，生成稀疏模型。S222. For space-time matrix and sparse function base library/> , through sparse regression, a sparse model is generated.
S223、根据序贯最小二乘回归法，确定稀疏解，根据稀疏模型以及稀疏解/>，建立非线性系统动力学稀疏识别算法SINDy模型。S223. Determine the sparse solution according to the sequential least squares regression method , according to the sparse model and sparse solution/> , establish the SINDy model of the nonlinear system dynamics sparse identification algorithm.
可选地，输入模块，进一步用于：Optionally, input modules are further used to:
对预处理后的历史数据，按照空间序列进行时空转换重采样，建立时空矩阵，如下（2）所示：For the preprocessed historical data, perform spatiotemporal conversion and resampling according to the spatial sequence to establish a spatiotemporal matrix. , as shown in (2) below:
（2） (2)
其中，表示空间序列，/>表示预处理后的历史数据，/>表示矩阵转置。in, Represents a spatial sequence,/> Represents preprocessed historical data,/> Represents matrix transpose.
可选地，稀疏模型，如下（3）所示：Optionally, a sparse model , as shown in (3) below:
（3） (3)
其中，表示稀疏函数基库，/>表示稀疏解。in, Represents a sparse function base library, /> represents a sparse solution.
可选地，非线性系统动力学稀疏识别算法SINDy模型，如下（4）所示：Optionally, nonlinear system dynamics sparse identification algorithm SINDy model , as shown in (4) below:
（4） (4)
其中，表示定义系统运动方程的动态约束，/>表示稀疏解，/>表示矩阵转置， 表示/>元素符号函数的向量。in, Represents dynamic constraints that define the system's equations of motion,/> Represents a sparse solution,/> represents the matrix transpose, Express/> Vector of element-wise symbolic functions.
可选地，输入模块，进一步用于：Optionally, input modules are further used to:
S231、设定由预处理后的历史数据所组成的矩阵，矩阵中的元素包括随时间演变的第一时序数据向量以及与第一时序数据向量/>存在线性关系的第二时序数据向量/>，且/>，其中，/>为状态矩阵。S231. Set a matrix composed of preprocessed historical data. The elements in the matrix include the first time series data vector that evolves over time. and with the first timing data vector/> The second time series data vector with a linear relationship/> , and/> , where,/> is the state matrix.
S232、对第一时序数据向量进行简化的奇异值分解SVD，得到状态矩阵/>的模态。S232. For the first time series data vector Perform simplified singular value decomposition SVD to obtain the state matrix/> modal.
S233、根据简化的奇异值分解SVD以及状态矩阵的模态，得到任意时间点的历史数据。S233. Based on simplified singular value decomposition SVD and state matrix mode to obtain historical data at any point in time.
可选地，任意时间点的历史数据，如下（5）所示：Optionally, historical data at any point in time , as shown in (5) below:
（5） (5)
其中，表示模态的总数，/>表示本征向量，/>表示时间点t下的稀疏旋钮，/>表示各模式的模态振幅。in, Indicates the total number of modalities,/> represents the eigenvector,/> Represents the sparse knob at time point t, /> Represents the modal amplitude of each mode.
一方面，提供了一种电子设备，所述电子设备包括处理器和存储器，所述存储器中存储有至少一条指令，所述至少一条指令由所述处理器加载并执行以实现上述热轧数字孪生带钢横截面形状实时预测方法。In one aspect, an electronic device is provided. The electronic device includes a processor and a memory. At least one instruction is stored in the memory. The at least one instruction is loaded and executed by the processor to realize the above hot-rolled digital twin. Real-time prediction method of strip cross-sectional shape.
一方面，提供了一种计算机可读存储介质，所述存储介质中存储有至少一条指令，所述至少一条指令由处理器加载并执行以实现上述热轧数字孪生带钢横截面形状实时预测方法。On the one hand, a computer-readable storage medium is provided. At least one instruction is stored in the storage medium. The at least one instruction is loaded and executed by a processor to implement the above-mentioned real-time prediction method of hot-rolled digital twin strip cross-sectional shape. .
上述技术方案，与现有技术相比至少具有如下有益效果：Compared with the existing technology, the above technical solution has at least the following beneficial effects:
上述方案，为了获取板形设定的最优工艺参数，本发明建立了高精度热轧带钢横截面形状预测模型，开发热连轧数字孪生生产线，实现任意工艺参数设定下的带钢虚拟热轧，并通过带钢数字孪生模型实时反映其带钢横断面形状。基于本方法及生产现场的带钢横断面形状控制精度要求，能够实现工艺设定参数的迭代优化，避免工艺参数设定不合理导致的热轧产品缺陷和降级现象的产生，降低钢铁企业的生产成本。In the above scheme, in order to obtain the optimal process parameters for plate shape setting, the present invention establishes a high-precision hot-rolled strip cross-sectional shape prediction model, develops a hot-rolling digital twin production line, and realizes virtual strip steel under arbitrary process parameter settings. Hot rolling, and the strip cross-sectional shape is reflected in real time through the strip digital twin model. Based on this method and the strip cross-sectional shape control accuracy requirements at the production site, iterative optimization of process setting parameters can be achieved, avoiding the occurrence of hot-rolled product defects and degradation caused by unreasonable process parameter settings, and reducing the production of steel enterprises. cost.
附图说明Description of the drawings
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without exerting creative efforts.
图1是本发明实施例提供的热轧数字孪生带钢横截面形状实时预测方法流程示意图；Figure 1 is a schematic flowchart of the real-time prediction method for cross-sectional shape of hot-rolled digital twin strip provided by an embodiment of the present invention;
图2是本发明实施例提供的序贯阈值最小二乘回归步骤图；Figure 2 is a step diagram of sequential threshold least squares regression provided by an embodiment of the present invention;
图3是本发明实施例提供的带钢横截面实际轧制结果；Figure 3 is the actual rolling result of the strip cross section provided by the embodiment of the present invention;
图4是本发明实施例提供的SINDy预测结果；Figure 4 is the SINDy prediction result provided by the embodiment of the present invention;
图5是本发明实施例提供的DMD-SINDy预测结果；Figure 5 is the DMD-SINDy prediction result provided by the embodiment of the present invention;
图6是本发明实施例提供的热轧数字孪生带钢横截面形状实时预测装置框图；Figure 6 is a block diagram of a real-time prediction device for hot-rolled digital twin strip cross-sectional shape provided by an embodiment of the present invention;
图7是本发明实施例提供的一种电子设备的结构示意图。FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention.
具体实施方式Detailed ways
为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例的附图，对本发明实施例的技术方案进行清楚、完整地描述。显然，所描述的实施例是本发明的一部分实施例，而不是全部的实施例。基于所描述的本发明的实施例，本领域普通技术人员在无需创造性劳动的前提下所获得的所有其他实施例，都属于本发明保护的范围。In order to make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the described embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.
如图1所示，本发明实施例提供了一种热轧数字孪生带钢横截面形状实时预测方法，该方法可以由电子设备实现。如图1所示的热轧数字孪生带钢横截面形状实时预测方法流程图，该方法的处理流程可以包括如下的步骤：As shown in Figure 1, an embodiment of the present invention provides a real-time prediction method for the cross-sectional shape of a hot-rolled digital twin strip, which can be implemented by electronic equipment. As shown in Figure 1, the flow chart of the real-time prediction method for the cross-sectional shape of the hot-rolled digital twin strip is shown. The processing flow of this method can include the following steps:
S1、构建热连轧数字孪生产线，获取热连轧数字孪生产线的设定参数。S1. Construct a hot rolling digital twin production line and obtain the setting parameters of the hot rolling digital twin production line.
一种可行的实施方式中，构建高精度热连轧数字孪生产线，并提供工艺过程参数设定接口。In a feasible implementation, a high-precision hot rolling digital twin production line is constructed and a process parameter setting interface is provided.
S2、采用DMD（Dynamic Mode Decomposition，动态模态分解）算法，优化SINDy（Sparse Identification of Nonlinear Dynamics，非线性系统动力学稀疏识别算法）模型，建立带钢横截面形状预测模型DMD-SINDy。S2. Use the DMD (Dynamic Mode Decomposition, dynamic mode decomposition) algorithm to optimize the SINDy (Sparse Identification of Nonlinear Dynamics, nonlinear system dynamics sparse identification algorithm) model, and establish the strip cross-section shape prediction model DMD-SINDy.
可选地，上述步骤S2可以包括如下步骤S21-23：Optionally, the above step S2 may include the following steps S21-23:
S21、收集热连轧生产过程的历史数据，对历史数据进行预处理。S21. Collect historical data of the hot continuous rolling production process and preprocess the historical data.
其中，历史数据包括实测轧制力、弯辊力、两侧轧制力弯辊力差以及凸度实测值。Among them, the historical data includes the measured rolling force, roll bending force, rolling force and roll bending force difference on both sides, and measured crown value.
进一步地，采集的生产数据无法直接用于建模，需要进行一定的预处理去除数据，通过对一维列向量样本数据进行线性变换，使数据分布至[0,1]范围内。Furthermore, the collected production data cannot be directly used for modeling. Certain preprocessing is required to remove the data. The one-dimensional column vector sample data is linearly transformed to distribute the data to the range [0,1].
具体地，对历史数据进行线性变换，如下式（1）所示：Specifically, linear transformation is performed on the historical data, as shown in the following equation (1):
（1） (1)
其中，表示线性变换后的数据，/>表示线性变换前的数据，/>表示数据样本中的最大数据，/>表示数据样本中的最小数据，/>表示历史数据数量。in, Represents the data after linear transformation,/> Represents data before linear transformation,/> Represents the maximum data in the data sample,/> Represents the minimum data in the data sample,/> Indicates the amount of historical data.
S22、建立非线性系统动力学稀疏识别算法SINDy模型。S22. Establish the SINDy model of the nonlinear system dynamics sparse identification algorithm.
可选地，上述步骤S22可以包括如下步骤S221-S223：Optionally, the above step S22 may include the following steps S221-S223:
S221、根据预处理后的历史数据，构建时空矩阵和稀疏函数基库/>。S221. Construct a space-time matrix based on the preprocessed historical data. and sparse function base library/> .
其中，构建时空矩阵，包括：Among them, construct the space-time matrix ,include:
采集时间序列的数据集并对数据按空间序列进行时空转换重采样，建立时空矩阵/>，如下（2）所示：Collect time series data sets and sequence the data spatially Perform spatio-temporal conversion resampling and establish a spatio-temporal matrix/> , as shown in (2) below:
（2） (2)
其中，表示带钢长度。in, Indicates the strip length.
进一步地，构建稀疏函数基库，它由x列的多个候选函数组成，如下（3）所示：Further, build a sparse function base library , which consists of multiple candidate functions of column x, as shown in (3) below:
（3） (3)
其中，以为例，表示在状态变量/>中的二次非线性项，具体如下：Among them, with For example, expressed in the state variable/> The quadratic nonlinear terms in are as follows:
（4） (4)
其中，表示空间序列上的数据。in, Represents data on a spatial sequence.
S222、根据从现场收集的实时生产数据，构成时空矩阵和输入变量矩阵/>。矩阵组合成字典矩阵/>，经过稀疏回归，产生一个稀疏模型：S222. Construct a space-time matrix based on real-time production data collected from the site. and input variable matrix/> . matrix Combined into dictionary matrix/> , after sparse regression, a sparse model is generated:
（5） (5)
其中，表示稀疏函数基库，/>表示稀疏解，/>中的每一列向量/>表示系数的稀疏向量。如果确定了/>，每行控制方程的模型可按如下方式构建：in, Represents a sparse function base library, /> Represents a sparse solution,/> Each column vector in/> A sparse vector representing coefficients. If confirmed/> , the model of each row of control equations can be constructed as follows:
（6） (6)
其中，是x元素符号函数的向量，不同于数据矩阵/>。整体模型就可以表示如下：in, is a vector of x-element symbolic functions, different from the data matrix/> . The overall model can be expressed as follows:
（7） (7)
S223、如图2所示，选择序贯最小二乘回归法确定回归问题的稀疏解，将低于阈值的权重强制设为0，然后用剩下的特征再做最小二乘，迭代若干次，进而求得最终的回归解。S223. As shown in Figure 2, select the sequential least squares regression method to determine the sparse solution to the regression problem. , force the weights below the threshold to 0, and then use the remaining features to perform least squares and iterate several times to obtain the final regression solution.
S23、采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，得到带钢横截面形状预测模型DMD-SINDy。S23. Use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm, and obtain the strip cross-sectional shape prediction model DMD-SINDy.
可选地，上述步骤S23可以包括如下步骤S231- S233：Optionally, the above step S23 may include the following steps S231-S233:
S231、设定由预处理后的历史数据所组成的矩阵，矩阵中的元素包括随时间演变的第一时序数据向量以及与第一时序数据向量/>存在线性关系的第二时序数据向量/>，且/>，其中，/>为状态矩阵。S231. Set a matrix composed of preprocessed historical data. The elements in the matrix include the first time series data vector that evolves over time. and with the first timing data vector/> The second time series data vector with a linear relationship/> , and/> , where,/> is the state matrix.
一种可行的实施方式中，假设实际热轧过程的量测数据所组成的矩阵为：In a feasible implementation, it is assumed that the matrix composed of the measurement data of the actual hot rolling process is:
（8） (8)
令系统的演化规律用线性关系描述：Let the evolution law of the system be described by a linear relationship:
（9） (9)
两组随时间演变的时序数据向量分别为：The two sets of time series data vectors evolving over time are:
（10） (10)
（11） (11)
其中和/>的维数均为/>：in and/> The dimensions of are all/> :
（12） (12)
对进行简化的SVD（Singular Value Decomposition，奇异值分解），有：right Simplified SVD (Singular Value Decomposition, singular value decomposition), there are:
（13） (13)
其中，为/>的数据矩阵；/>为/>的对角阵；/>为/>的正交阵。令，可得：in, for/> data matrix;/> for/> diagonal array;/> for/> of orthogonal array. make ,Available:
（14） (14)
则的本征向量为/>，并且满足：but The eigenvector of is/> , and satisfy:
（15） (15)
其中，表示稀疏旋钮。in, Indicates the sparse knob.
S232、对第一时序数据向量进行简化的奇异值分解SVD，得到原系统状态矩阵/>的模态/>为：S232. For the first time series data vector Perform simplified singular value decomposition SVD to obtain the original system state matrix/> Modal/> for:
（16） (16)
S233、基于上述动态模式分解过程，任意时间点的采样数据可表示为：S233. Based on the above dynamic mode decomposition process, the sampling data at any time point can be expressed as:
（17） (17)
其中，表示模态的总数，/>表示本征向量，/>表示时间点t下的稀疏旋钮，/>表示各模式的模态振幅。in, Indicates the total number of modalities,/> represents the eigenvector,/> Represents the sparse knob at time point t, /> Represents the modal amplitude of each mode.
S3、根据设定参数以及带钢横截面形状预测模型，得到热连轧数字孪生产线的带钢横截面形状预测结果。S3. According to the set parameters and the strip cross-sectional shape prediction model, obtain the strip cross-sectional shape prediction results of the hot continuous rolling digital twin production line.
一种可行的实施方式中，将数字孪生产线的人工设定参数输入至带钢横截面形状预测模型，得到带钢数字孪生模型的横截面形状。In a feasible implementation, the manually set parameters of the digital twin production line are input into the strip cross-sectional shape prediction model to obtain the cross-sectional shape of the strip digital twin model.
进一步地，对热轧数字孪生带钢横截面形状预测精度进行评价。Furthermore, the prediction accuracy of the cross-sectional shape of the hot-rolled digital twin strip was evaluated.
其中，热轧数字孪生带钢横截面形状预测精度评价指标具体包括：Among them, the evaluation indicators of the cross-sectional shape prediction accuracy of hot-rolled digital twin strips include:
（1）MAE（Mean Absolute Error，平均绝对误差）：(1) MAE (Mean Absolute Error, mean absolute error):
（18） (18)
（2）RMSE（Mean Squa re Error，均方根误差）：(2) RMSE (Mean Square Error, root mean square error):
（19） (19)
其中，为样本数量；/>为期望值；/>为模型预测值。其中MAE和RMSE越小，说明模型的预测精度越高。in, is the sample number;/> is the expected value;/> is the predicted value of the model. The smaller the MAE and RMSE are, the higher the prediction accuracy of the model is.
以某2250热连轧厂为实施案例，该产线产品带钢厚度范围1.2~25.4mm，宽度规格范围800~2130mm。末机架出口处的凸度检测仪布置有60个测量通道，通过每个测量通道内获得的多个带钢厚度值，可以组合构成带钢的横断面轮廓。本发明方法实施如下：Taking a 2250 hot continuous rolling mill as an implementation case, the strip thickness range of this production line ranges from 1.2 to 25.4mm, and the width specification ranges from 800 to 2130mm. The crown detector at the exit of the final frame is equipped with 60 measurement channels. Through multiple strip thickness values obtained in each measurement channel, the cross-sectional profile of the strip can be combined. The method of the present invention is implemented as follows:
步骤S1：收集的热轧生产过程历史数据参数项如表1所示，具体数值如表2所示。并对表2中的数据进行预处理。Step S1: The collected historical data parameters of the hot rolling production process are shown in Table 1, and the specific values are shown in Table 2. And preprocess the data in Table 2.
表1Table 1
表2Table 2
步骤S2：采用动态模态分解优化非线性系统动力学稀疏识别算法，建立带钢横截面形状预测模型DMD-SINDy。Step S2: Use dynamic mode decomposition to optimize the nonlinear system dynamics sparse identification algorithm and establish the strip cross-sectional shape prediction model DMD-SINDy.
步骤S3：构建高精度热连轧数字孪生产线，并提供工艺过程参数设定接口。Step S3: Construct a high-precision hot rolling digital twin production line and provide a process parameter setting interface.
步骤S4：以某卷钢为例，将数字孪生产线的人工设定参数输入至DMD-SINDy模型，预测得到带钢数字孪生模型的横截面形状。为了更好的对比本方法的有效性，选择相同工艺参数轧制后的实测板形和SINDy预测的板形进行对比，具体结果如图3-5所示；Step S4: Taking a certain coil of steel as an example, input the manually set parameters of the digital twin production line into the DMD-SINDy model, and predict the cross-sectional shape of the digital twin model of the strip steel. In order to better compare the effectiveness of this method, the measured plate shape after rolling with the same process parameters and the plate shape predicted by SINDy were selected for comparison. The specific results are shown in Figure 3-5;
步骤S5：对热轧数字孪生带钢横截面形状预测精度进行评价，评价结果如表3所示：Step S5: Evaluate the prediction accuracy of the cross-sectional shape of the hot-rolled digital twin strip. The evaluation results are shown in Table 3:
表3table 3
通过图3-5和表3可知，本发明方法预测精度最高，能够满足生产现场的使用要求。依托本发明搭建的热连轧数字孪生平台，可进行虚拟仿真轧制，得到不同热轧工艺参数设定下的带钢横截面形状。这对于实现轧制工艺设定参数的迭代优化，避免工艺参数设定不合理导致的热轧产品缺陷和降级现象的产生，降低钢铁企业的生产成本具有重要意义。It can be seen from Figures 3-5 and Table 3 that the method of the present invention has the highest prediction accuracy and can meet the requirements of the production site. Relying on the hot continuous rolling digital twin platform built by the present invention, virtual simulation rolling can be performed to obtain the strip cross-sectional shape under different hot rolling process parameter settings. This is of great significance for realizing iterative optimization of rolling process setting parameters, avoiding the occurrence of hot-rolled product defects and degradation caused by unreasonable process parameter setting, and reducing the production costs of steel enterprises.
本发明实施例中，为了获取板形设定的最优工艺参数，本发明建立了高精度热轧带钢横截面形状预测模型，开发热连轧数字孪生生产线，实现任意工艺参数设定下的带钢虚拟热轧，并通过带钢数字孪生模型实时反映其带钢横断面形状。基于本方法及生产现场的带钢横断面形状控制精度要求，能够实现工艺设定参数的迭代优化，避免工艺参数设定不合理导致的热轧产品缺陷和降级现象的产生，降低钢铁企业的生产成本。In the embodiment of the present invention, in order to obtain the optimal process parameters for plate shape setting, the present invention establishes a high-precision hot-rolled strip cross-sectional shape prediction model, develops a hot-rolling digital twin production line, and realizes the optimal process parameters under arbitrary process parameter settings. The strip is virtually hot-rolled, and its strip cross-sectional shape is reflected in real time through the strip digital twin model. Based on this method and the strip cross-sectional shape control accuracy requirements at the production site, iterative optimization of process setting parameters can be achieved, avoiding the occurrence of hot-rolled product defects and degradation caused by unreasonable process parameter settings, and reducing the production of steel enterprises. cost.
如图6所示，本发明实施例提供了一种热轧数字孪生带钢横截面形状实时预测装置600，该装置600应用于实现热轧数字孪生带钢横截面形状实时预测方法，该装置600包括：As shown in Figure 6, an embodiment of the present invention provides a real-time prediction device 600 for the cross-sectional shape of a hot-rolled digital twin strip. The device 600 is used to implement a real-time prediction method for the cross-sectional shape of a hot-rolled digital twin strip. The device 600 include:
获取模块610，用于构建热连轧数字孪生产线，获取热连轧数字孪生产线的设定参数。The acquisition module 610 is used to construct a hot rolling digital twin production line and obtain the setting parameters of the hot rolling digital twin production line.
输入模块620，用于采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，建立带钢横截面形状预测模型DMD-SINDy。The input module 620 is used to use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm and establish the strip cross-sectional shape prediction model DMD-SINDy.
输出模块630，用于根据设定参数以及带钢横截面形状预测模型，得到热连轧数字孪生产线的带钢横截面形状预测结果。The output module 630 is used to obtain the strip cross-sectional shape prediction results of the hot continuous rolling digital twin production line based on the set parameters and the strip cross-sectional shape prediction model.
可选地，输入模块620，进一步用于：Optionally, the input module 620 is further used for:
S21、收集热连轧生产过程的历史数据，对历史数据进行预处理。S21. Collect historical data of the hot continuous rolling production process and preprocess the historical data.
其中，历史数据包括实测轧制力、弯辊力、两侧轧制力弯辊力差以及凸度实测值。Among them, the historical data includes the measured rolling force, roll bending force, rolling force and roll bending force difference on both sides, and measured crown value.
S22、建立非线性系统动力学稀疏识别算法SINDy模型。S22. Establish the SINDy model of the nonlinear system dynamics sparse identification algorithm.
S23、采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，得到带钢横截面形状预测模型DMD-SINDy。S23. Use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm, and obtain the strip cross-sectional shape prediction model DMD-SINDy.
可选地，输入模块620，进一步用于：Optionally, the input module 620 is further used for:
对历史数据进行线性变换，如下式（1）所示：Perform linear transformation on historical data, as shown in the following equation (1):
（1） (1)
其中，表示线性变换后的数据，/>表示线性变换前的数据，/>表示历史数据数量。in, Represents the data after linear transformation,/> Represents data before linear transformation,/> Indicates the amount of historical data.
可选地，输入模块620，进一步用于：Optionally, the input module 620 is further used for:
S221、根据预处理后的历史数据，构建时空矩阵和稀疏函数基库/>。S221. Construct a space-time matrix based on the preprocessed historical data. and sparse function base library/> .
S222、对时空矩阵和稀疏函数基库/>，经过稀疏回归，生成稀疏模型。S222. For space-time matrix and sparse function base library/> , through sparse regression, a sparse model is generated.
S223、根据序贯最小二乘回归法，确定稀疏解，根据稀疏模型以及稀疏解/>，建立非线性系统动力学稀疏识别算法SINDy模型。S223. Determine the sparse solution according to the sequential least squares regression method , according to the sparse model and sparse solution/> , establish the SINDy model of the nonlinear system dynamics sparse identification algorithm.
可选地，输入模块620，进一步用于：Optionally, the input module 620 is further used for:
对预处理后的历史数据，按照空间序列进行时空转换重采样，建立时空矩阵，如下（2）所示：For the preprocessed historical data, perform spatiotemporal conversion and resampling according to the spatial sequence to establish a spatiotemporal matrix. , as shown in (2) below:
（2） (2)
其中，表示空间序列，/>表示预处理后的历史数据，/>表示矩阵转置。in, Represents a spatial sequence,/> Represents preprocessed historical data,/> Represents matrix transpose.
可选地，稀疏模型，如下（3）所示：Optionally, a sparse model , as shown in (3) below:
（3） (3)
其中，表示稀疏函数基库，/>表示稀疏解。in, Represents a sparse function base library, /> represents a sparse solution.
可选地，非线性系统动力学稀疏识别算法SINDy模型，如下（4）所示：Optionally, nonlinear system dynamics sparse identification algorithm SINDy model , as shown in (4) below:
（4） (4)
其中，表示定义系统运动方程的动态约束，/>表示稀疏解，/>表示矩阵转置， 表示/>元素符号函数的向量。in, Represents dynamic constraints that define the system's equations of motion,/> Represents a sparse solution,/> represents the matrix transpose, Express/> Vector of element-wise symbolic functions.
可选地，输入模块620，进一步用于：Optionally, the input module 620 is further used for:
S231、设定由预处理后的历史数据所组成的矩阵，矩阵中的元素包括随时间演变的第一时序数据向量以及与第一时序数据向量/>存在线性关系的第二时序数据向量/>，且/>，其中，/>为状态矩阵。S231. Set a matrix composed of preprocessed historical data. The elements in the matrix include the first time series data vector that evolves over time. and with the first timing data vector/> The second time series data vector with a linear relationship/> , and/> , where,/> is the state matrix.
S232、对第一时序数据向量进行简化的奇异值分解SVD，得到状态矩阵/>的模态。S232. For the first time series data vector Perform simplified singular value decomposition SVD to obtain the state matrix/> modal.
S233、根据简化的奇异值分解SVD以及状态矩阵的模态，得到任意时间点的历史数据。S233. Based on simplified singular value decomposition SVD and state matrix mode to obtain historical data at any point in time.
可选地，任意时间点的历史数据，如下（5）所示：Optionally, historical data at any point in time , as shown in (5) below:
（5） (5)
其中，表示模态的总数，/>表示本征向量，/>表示时间点t下的稀疏旋钮，/>表示各模式的模态振幅。in, Indicates the total number of modalities,/> represents the eigenvector,/> Represents the sparse knob at time point t, /> Represents the modal amplitude of each mode.
本发明实施例中，为了获取板形设定的最优工艺参数，本发明建立了高精度热轧带钢横截面形状预测模型，开发热连轧数字孪生生产线，实现任意工艺参数设定下的带钢虚拟热轧，并通过带钢数字孪生模型实时反映其带钢横断面形状。基于本方法及生产现场的带钢横断面形状控制精度要求，能够实现工艺设定参数的迭代优化，避免工艺参数设定不合理导致的热轧产品缺陷和降级现象的产生，降低钢铁企业的生产成本。In the embodiment of the present invention, in order to obtain the optimal process parameters for plate shape setting, the present invention establishes a high-precision hot-rolled strip cross-sectional shape prediction model, develops a hot-rolling digital twin production line, and realizes the optimal process parameters under arbitrary process parameter settings. The strip is virtually hot-rolled, and its strip cross-sectional shape is reflected in real time through the strip digital twin model. Based on this method and the strip cross-sectional shape control accuracy requirements at the production site, iterative optimization of process setting parameters can be achieved, avoiding the occurrence of hot-rolled product defects and degradation caused by unreasonable process parameter settings, and reducing the production of steel enterprises. cost.
图7是本发明实施例提供的一种电子设备700的结构示意图，该电子设备700可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上处理器（centralprocessing units，CPU）701和一个或一个以上的存储器702，其中，存储器702中存储有至少一条指令，至少一条指令由处理器701加载并执行以实现下述热轧数字孪生带钢横截面形状实时预测方法：Figure 7 is a schematic structural diagram of an electronic device 700 provided by an embodiment of the present invention. The electronic device 700 may vary greatly due to different configurations or performance, and may include one or more processors (central processing units, CPU) 701 and one or more memories 702, wherein at least one instruction is stored in the memory 702, and at least one instruction is loaded and executed by the processor 701 to implement the following real-time prediction method of hot-rolled digital twin strip cross-sectional shape:
S1、构建热连轧数字孪生产线，获取热连轧数字孪生产线的设定参数。S1. Construct a hot rolling digital twin production line and obtain the setting parameters of the hot rolling digital twin production line.
S2、采用动态模态分解DMD算法，优化非线性系统动力学稀疏识别算法SINDy模型，建立带钢横截面形状预测模型DMD-SINDy。S2. Use the dynamic mode decomposition DMD algorithm to optimize the SINDy model of the nonlinear system dynamics sparse identification algorithm and establish the strip cross-sectional shape prediction model DMD-SINDy.
S3、根据设定参数以及带钢横截面形状预测模型，得到热连轧数字孪生产线的带钢横截面形状预测结果。S3. According to the set parameters and the strip cross-sectional shape prediction model, obtain the strip cross-sectional shape prediction results of the hot continuous rolling digital twin production line.
在示例性实施例中，还提供了一种计算机可读存储介质，例如包括指令的存储器，上述指令可由终端中的处理器执行以完成上述热轧数字孪生带钢横截面形状实时预测方法。例如，计算机可读存储介质可以是ROM、随机存取存储器（RAM）、CD-ROM、磁带、软盘和光数据存储设备等。In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including instructions, and the instructions can be executed by a processor in a terminal to complete the above-mentioned real-time prediction method of hot-rolled digital twin strip cross-sectional shape. For example, computer-readable storage media may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。Those of ordinary skill in the art can understand that all or part of the steps to implement the above embodiments can be completed by hardware, or can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable storage medium. The above-mentioned The storage media mentioned can be read-only memory, magnetic disks or optical disks, etc.
以上所述仅为本发明的较佳实施例，并不用以限制本发明，凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within the range.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311436735.4ACN117150832B (en) | 2023-11-01 | 2023-11-01 | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311436735.4ACN117150832B (en) | 2023-11-01 | 2023-11-01 | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel |
| Publication Number | Publication Date |
|---|---|
| CN117150832Atrue CN117150832A (en) | 2023-12-01 |
| CN117150832B CN117150832B (en) | 2024-02-23 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311436735.4AActiveCN117150832B (en) | 2023-11-01 | 2023-11-01 | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel |
| Country | Link |
|---|---|
| CN (1) | CN117150832B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118988999A (en)* | 2024-10-24 | 2024-11-22 | 承德建龙特殊钢有限公司 | Method and device for predicting rolling force of hot-rolled seamless steel tube based on digital twin |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343606A (en)* | 2021-07-01 | 2021-09-03 | 浙江大学 | Method for predicting full flow field from sparse sensor information based on compressed sensing reduced order model |
| CN113333474A (en)* | 2021-04-27 | 2021-09-03 | 山东钢铁股份有限公司 | Strip steel hot-rolled plate shape control method and system based on digital twinning |
| CN113504768A (en)* | 2021-08-05 | 2021-10-15 | 东华大学 | High-precision product digital twin computability method for assembly quality prediction |
| WO2022121923A1 (en)* | 2020-12-10 | 2022-06-16 | 东北大学 | Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium |
| CN114647904A (en)* | 2022-02-24 | 2022-06-21 | 北京工业大学 | An adaptive sparse graph learning bearing life prediction method based on digital twin dictionary |
| CN115034152A (en)* | 2022-05-17 | 2022-09-09 | 浙江大学 | A data-driven nonlinear reduced-order prediction method and device for fluid-structure coupled systems |
| CN115796002A (en)* | 2022-11-01 | 2023-03-14 | 北京科技大学 | Strip steel coil defect prediction and diagnosis method based on deep learning |
| CN116020879A (en)* | 2023-02-15 | 2023-04-28 | 北京科技大学 | Technological parameter-oriented strip steel hot continuous rolling space-time multi-scale process monitoring method and device |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022121923A1 (en)* | 2020-12-10 | 2022-06-16 | 东北大学 | Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium |
| CN113333474A (en)* | 2021-04-27 | 2021-09-03 | 山东钢铁股份有限公司 | Strip steel hot-rolled plate shape control method and system based on digital twinning |
| CN113343606A (en)* | 2021-07-01 | 2021-09-03 | 浙江大学 | Method for predicting full flow field from sparse sensor information based on compressed sensing reduced order model |
| CN113504768A (en)* | 2021-08-05 | 2021-10-15 | 东华大学 | High-precision product digital twin computability method for assembly quality prediction |
| CN114647904A (en)* | 2022-02-24 | 2022-06-21 | 北京工业大学 | An adaptive sparse graph learning bearing life prediction method based on digital twin dictionary |
| CN115034152A (en)* | 2022-05-17 | 2022-09-09 | 浙江大学 | A data-driven nonlinear reduced-order prediction method and device for fluid-structure coupled systems |
| CN115796002A (en)* | 2022-11-01 | 2023-03-14 | 北京科技大学 | Strip steel coil defect prediction and diagnosis method based on deep learning |
| CN116020879A (en)* | 2023-02-15 | 2023-04-28 | 北京科技大学 | Technological parameter-oriented strip steel hot continuous rolling space-time multi-scale process monitoring method and device |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118988999A (en)* | 2024-10-24 | 2024-11-22 | 承德建龙特殊钢有限公司 | Method and device for predicting rolling force of hot-rolled seamless steel tube based on digital twin |
| Publication number | Publication date |
|---|---|
| CN117150832B (en) | 2024-02-23 |
| Publication | Publication Date | Title |
|---|---|---|
| CN108510006B (en) | Enterprise power consumption analysis and prediction method based on data mining | |
| Bucak et al. | Incremental subspace learning via non-negative matrix factorization | |
| CN117150832B (en) | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel | |
| CN115081717B (en) | Rail transit passenger flow prediction method integrating attention mechanism and graph neural network | |
| CN111241657A (en) | Cold-rolled strip steel plate shape prediction method based on deep learning | |
| Zhao et al. | A novel paradigm of flatness prediction and optimization for strip tandem cold rolling by cloud-edge collaboration | |
| CN114798727B (en) | Multi-objective optimization-based specification self-adaptive rolling method and device and electronic equipment | |
| CN115600717B (en) | Sub-industry power load prediction method, equipment and storage medium based on federal learning | |
| CN103778466B (en) | Converter steel-making temperature modeling forecasting method based on vector error and system thereof | |
| Han et al. | Prediction and control of profile for silicon steel strip in the whole tandem cold rolling based on PSO-BP algorithm | |
| WO2022111002A1 (en) | Method and apparatus for training neural network, and computer readable storage medium | |
| Li et al. | Modeling and validation of bending force for 6-high tandem cold rolling mill based on machine learning models | |
| CN113836618A (en) | Temporary steel buckle tower disassembly error adjustment control method based on support vector machine | |
| Zhao et al. | Predicting flatness of strip tandem cold rolling using a general regression neural network optimized by differential evolution algorithm | |
| Yuan et al. | Research on quarter wave shape control of cold rolled high strength steel based on BP-MOPSO | |
| CN117856254A (en) | Power grid direct control load control method based on neural network | |
| CN117216526A (en) | Photovoltaic output prediction method and system based on artificial intelligence | |
| Li et al. | An industrial IoT-based deformation resistance prediction and thickness control method of cold-rolled strip in steel production systems | |
| CN118988997B (en) | Cross-process feedforward control method suitable for strip steel plate shape of continuous annealing leveling unit | |
| Guo | A block bootstrap for quasi-likelihood in sparse functional data | |
| CN119513696A (en) | A method, system, device and storage medium for reducing the order of power transmission and transformation model | |
| CN110222825A (en) | A kind of finished cement specific surface area prediction technique and system | |
| CN114595890A (en) | Prediction method and system of ship spare parts demand based on BP-SVR combined model | |
| CN117891214B (en) | Method and device for monitoring and fault diagnosis of hot continuous rolling process under Yun Bianduan cooperation | |
| CN110773580B (en) | Method and system for calculating rolling force based on XGboost algorithm |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |