A kpi related fault diagnosis method and device for multi-mode manufacturing process

Abstract

The present application relates to the technical field of control and monitoring of industrial production process, and particularly relates to a KPI related fault diagnosis method and device for multi-mode manufacturing process. The method comprises: using a supervising variant SMGU of MGU for KPI related feature extraction. Process variables and KPIs are used as inputs of SMGU, and SMGU learns potential hidden states to effectively maintain their connection with KPI information. A sparse variant SBLS of BLS is used for fault diagnosis, thereby reducing network structure redundancy and avoiding model overfitting. The present application solves the problems of poor universality and low diagnosis accuracy of traditional deep learning based fault diagnosis methods, and improves the timeliness and accuracy of fault diagnosis. The fault diagnosis method has good scalability, and when new training data is added, the complete training process does not need to be repeated, and the method can be effectively applied to multi-mode manufacturing processes such as hot rolling of strip steel.

Description

Technical Field

[0001] This invention relates to the field of industrial production process control and monitoring technology, and in particular to a method and device for diagnosing KPI-related faults in multi-mode manufacturing processes. Background Technology

[0002] With the rapid advancement of computer and sensor technologies, manufacturing processes are increasingly incorporating large-scale information integration and strong interaction. This trend stems from the need to adapt to frequent adjustments in product specifications, changes in raw material quality, and other uncertain external factors. Compared to single-mode manufacturing processes, multi-mode manufacturing processes present significant challenges due to their increased production scale and complexity. Fault diagnosis technology for multi-mode manufacturing processes is a key technology for improving equipment safety, reliability, and reducing accident risks.

[0003] In modern industrial production, safety and stability are paramount. Accidents or malfunctions in the industrial production process can negatively impact product quality and production efficiency, and may even cause personal injury or death. Therefore, developing a safe and efficient fault diagnosis technology is of great significance.

[0004] Multimodal fault diagnosis can be categorized into methods based on multivariate statistical analysis, machine learning, and deep learning. Early research primarily focused on multivariate statistical analysis and machine learning methods, including Mode Transition Probability (MTP), Principal Component Analysis (PCA), K-Nearest Neighbors (KNN), and K-means. However, statistical and machine learning-based methods have limited effectiveness for high-dimensional, nonlinear, and non-stationary data, failing to meet practical industrial needs. Compared to other fields, the application of deep learning methods in multimodal fault diagnosis is relatively limited; related research employs transfer learning-based methods to handle multimodal manufacturing processes. KPI-related fault diagnosis includes methods such as Partial Least Squares (APLS) and autoencoders.

[0005] Most existing deep learning-based fault diagnosis methods for multi-modal manufacturing processes employ static deep networks such as autoencoders. However, multi-modal manufacturing processes are inherently dynamic, and existing methods struggle to fully extract the dynamic features of the data. Furthermore, they lack a focus on extracting KPI-related features. Simultaneously, the large network structure of deep learning models reduces their adaptability and flexibility, leading to overfitting and consequently lower fault diagnosis accuracy. Summary of the Invention

[0006] To address the shortcomings of existing methods in extracting key KPI-related features, and the limitations of deep learning models due to their large network structures which reduce adaptability and flexibility and lead to overfitting and low fault diagnosis accuracy, this invention provides a KPI-related fault diagnosis method and apparatus for multi-mode manufacturing processes. The technical solution is as follows:

[0007] On the one hand, a method for diagnosing KPI-related faults in multi-mode manufacturing processes is provided, characterized in that the method includes:

[0008] S1. Obtain process variables and key performance indicators (KPIs) data retained during the hot rolling process of strip steel, and establish a historical sample database;

[0009] S2. Construct a supervised minimum gate unit (SMGU), train the SMGU based on a historical sample database, and use the SMGU to extract dynamic features related to the key performance indicator (KPI) data.

[0010] S3. Construct a Sparse Width Learning System (SBLS) and input the dynamic features related to the extracted Key Performance Indicator (KPI) data into the Sparse Width Learning System (SBLS) for fine-grained fault diagnosis training.

[0011] S4. Based on the trained Supervised Minimum Gated Unit (SMGU) and Sparse Width Learning System (SBLS), obtain the KPI-related fault diagnosis structure based on SMGU-SBLS.

[0012] S5. Obtain the fault data to be detected, input the fault data to be detected into the KPI-related fault diagnosis structure based on SMGU-SBLS, and obtain the fault diagnosis result.

[0013] Optionally, in S1, process variables include: various data sets generated during the hot rolling process of strip steel when producing steel plates of different steel grades and thicknesses;

[0014] Key performance indicators (KPIs) include the width, crown, thickness, and flatness of the strip exiting the finishing mill.

[0015] Optionally, in S2, a supervised minimum gated unit (SMGU) is constructed and trained based on a historical sample database; the SMGU is then used to extract dynamic features related to the key performance indicator (KPI) data, including:

[0016] Construct a supervised minimum gated unit (SMGU);

[0017] Based on key performance indicator (KPI) data and process variables, update rules for supervised minimum gated unit (SMGU) are constructed to obtain the network parameters of the supervised minimum gated unit (SMGU).

[0018] The output of the supervised minimum gating unit (SMGU) is used as a dynamic feature related to the key performance indicator (KPI) data.

[0019] Optionally, S3, a sparse width learning system (SBLS) is constructed, and the dynamic features related to the extracted key performance indicator (KPI) data are input into the SBLS for fine-grained fault diagnosis training, including:

[0020] Constructing a Sparse Width Learning System (SBLS);

[0021] Based on the dynamic characteristics of the extracted key performance indicator (KPI) data, augmented Lagrange multipliers are used to perform unconstrained optimization of the objective function of SBLS. The unconstrained optimization includes solving for the intermediate variables and parameters of the objective function using block coordinate descent (BCD).

[0022] Based on the dynamic characteristics related to key performance indicator (KPI) data, and combined with intermediate variables and parameters of the objective function, fault diagnosis results are obtained.

[0023] Optionally, the fault diagnosis results include:

[0024] Fault diagnosis results The output of SBLS is calculated as follows:

[0025]

[0026] in, Here is the weight matrix of SBLS. The weights represent the final weights obtained after several iterations, and A represents the actual input matrix calculated from the dynamic features F related to the key performance indicator (KPI) data.

[0027] Optionally, in S5, the fault data to be detected is acquired, and the fault data is input into the SMGU-SBLS-based KPI-related fault diagnosis structure to obtain the fault diagnosis results, including:

[0028] Acquire the fault data to be detected, and the fault type of the fault data;

[0029] The SMGU-SBLS-based KPI-related fault diagnosis structure is used to extract features from the fault data to be detected, and to calculate the input matrix and weight matrix.

[0030] Fault diagnosis results are obtained based on the input matrix and weight matrix.

[0031] On the other hand, a KPI-related fault diagnosis device for multi-mode manufacturing processes is provided. This device is applied to a KPI-related fault diagnosis method for multi-mode manufacturing processes. The device includes:

[0032] The data acquisition module is used to acquire process variables and key performance indicators (KPIs) data retained during the hot rolling of strip steel and to establish a historical sample database.

[0033] The feature extraction module is used to construct a supervised minimum gate unit (SMGU), train the SMGU based on a historical sample database, and extract dynamic features related to key performance indicator (KPI) data using the SMGU.

[0034] The SBLS building module is used to build the Sparse Width Learning System (SBLS). The dynamic features related to the extracted key performance indicator (KPI) data are input into the Sparse Width Learning System (SBLS) for fine-grained fault diagnosis training.

[0035] The SMGU-SBLS structure building module is used to obtain a KPI-related fault diagnosis structure based on SMGU-SBLS based on the trained supervised minimum gate unit SMGU and sparse width learning system SBLS.

[0036] The fault diagnosis module is used to acquire the fault data to be detected, input the fault data to be detected into the KPI-related fault diagnosis structure based on SMGU-SBLS, and obtain the fault diagnosis results.

[0037] Optionally, process variables include: various data sets generated during the hot rolling process of strip steel when producing steel plates of different steel grades and thicknesses;

[0038] Key performance indicators (KPIs) include the width, crown, thickness, and flatness of the strip exiting the finishing mill.

[0039] On the other hand, a KPI-related fault diagnosis device for multi-mode manufacturing processes is provided. The KPI-related fault diagnosis device for multi-mode manufacturing processes includes: a processor; a memory, wherein computer-readable instructions are stored in the memory, and when the computer-readable instructions are executed by the processor, any one of the methods described above for KPI-related fault diagnosis of multi-mode manufacturing processes is implemented.

[0040] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored therein, the at least one instruction being loaded and executed by a processor to implement any of the above-described methods for KPI-related fault diagnosis of multi-mode manufacturing processes.

[0041] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0042] This invention presents a KPI-related dynamic feature extraction method based on SMGU: MGU significantly reduces training complexity and improves training speed by simplifying the gate structure. However, unsupervised MGU primarily learns dynamic features from process variables, which may contain information unrelated to KPIs. Supervised SMGU can fully extract KPI-related dynamic characteristics.

[0043] Fine-grained fault diagnosis based on SBLS: The large number of nodes in BLS increases the complexity of the network. Sparse BLS simplifies the structure of traditional BLS by removing redundant feature nodes and enhancing nodes, effectively avoiding overfitting of the model and enhancing the interpretability of the network. Attached Figure Description

[0044] 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a flowchart of a KPI-related fault diagnosis method for multi-mode manufacturing processes provided by an embodiment of the present invention;

[0046] Figure 2 This is a structural diagram of the SMGU provided in an embodiment of the present invention;

[0047] Figure 3 This is a diagram of the KPI-related fault diagnosis structure based on SMGU-SBLS provided in this embodiment of the invention;

[0048] Figure 4 This is a diagram illustrating the extended capabilities of SMGU-SBLS provided in this embodiment of the invention;

[0049] Figure 5 These are accuracy graphs for different training samples provided in the embodiments of the present invention;

[0050] Figure 6 This is a block diagram of a KPI-related fault diagnosis device for multi-mode manufacturing processes provided in an embodiment of the present invention;

[0051] Figure 7 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0052] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0053] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0054] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0055] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0056] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0057] This invention provides a KPI-related fault diagnosis method for multi-mode manufacturing processes. This method can be implemented by a KPI-related fault diagnosis device for multi-mode manufacturing processes, which can be a terminal or a server. Figure 1 The flowchart shown is for a KPI-related fault diagnosis method for multi-mode manufacturing processes. The processing flow of this method may include the following steps:

[0058] S1. Obtain process variables and key performance indicators (KPIs) data retained during the hot rolling process of strip steel, and establish a historical sample database;

[0059] In one feasible implementation, in S1, the process variables include: various data sets generated during the hot rolling process of strip steel when producing steel plates of different steel grades and different thicknesses;

[0060] Key performance indicators (KPIs) include the width, crown, thickness, and flatness of the strip exiting the finishing mill.

[0061] In one feasible implementation, Key Performance Indicators (KPIs) are an important concept in the manufacturing process, particularly those related to key parameters of production, product content, and critical equipment. These KPIs require close attention because they establish quantitative relationships between the performance of lower-level technology cycles / components and higher-level indicators such as product quality, production efficiency, energy consumption, and raw material consumption.

[0062] In one feasible implementation, the historical sample data obtained by the present invention consists of various data sets (normal and different faults) generated when steel plates of different steel grades and thicknesses were previously produced. The KPIs include the width, crown, thickness and flatness of the strip at the exit of the finishing mill, and the process variables include the roll gap, rolling force, rolling speed of stands 1-7 and the bending roll force of stands 2-7 (no bending roll force in stand 1), as shown in Table 1.

[0063] Table 1. Variables in the hot rolling process of strip steel

[0064]

[0065] S2. Construct a supervised minimum gate unit (SMGU), train the SMGU based on a historical sample database, and use the SMGU to extract dynamic features related to the key performance indicator (KPI) data.

[0066] In one feasible implementation, the minimal gated unit (MGU) is a type of gated unit in recurrent neural networks (RNNs). An MGU contains only one forget gate and is the smallest design among all gated hidden units. The MGU has comparable accuracy to a gated recurrent unit (GRU), but with a simpler structure, fewer parameters, and faster training speed.

[0067] In one feasible implementation, in S2, a supervised minimum gating unit (SMGU) is used to extract dynamic features related to key performance indicator (KPI) data. Compared with a conventional MGU, the SMGU uses KPI data as another input to guide the learning of the hidden state, thus the feature extraction is supervised learning. The method includes:

[0068] Based on key performance indicator (KPI) data, update rules for supervised minimum gated unit (SMGU) are constructed, and network parameters of the supervised minimum gated unit (SMGU) are obtained.

[0069] The update rules for the supervised minimum gating unit are summarized as follows:

[0070]

[0071] In the formula, for Forget the door at all times for Time-process data, for KPI value at any time for Candidate hidden state at any time. , and These are the training weights for the forget gate. , , for Weight, the first letter of the subscript of the weight indicates that the weight belongs to or The second letter indicates that the weight belongs to or wait. and They are respectively and The activation function.

[0072] In one feasible implementation, S2 further includes:

[0073] The network parameters of the supervised minimum gated unit (SMGU) are fine-tuned using the backpropagation algorithm:

[0074]

[0075] In the formula, for or , For the number of variables, It is the learning rate. Among them... The loss function is shown in the SMGU structure diagram. Figure 2 As shown.

[0076] S3. Construct a Sparse Width Learning System (SBLS) and input the dynamic features related to the extracted Key Performance Indicator (KPI) data into the Sparse Width Learning System (SBLS) for fine-grained fault diagnosis training.

[0077] In one feasible implementation, the broad learning system (BLS) is an efficient incremental learning system comprising a three-layer structure: an input layer, hidden layers (feature nodes and augmentation nodes), and an output layer. Unlike deep learning algorithms that train models through gradient descent and backpropagation, BLS uses the ridge regression learning algorithm to solve the pseudo-inverse between hidden layer nodes and the output layer to obtain model parameters, thereby solving regression and classification problems.

[0078] In one feasible implementation, S3, based on the dynamic features related to the extracted key performance indicator (KPI) data, a sparse width learning system is used for fine-grained fault diagnosis. Compared with conventional width learning, this system enhances the interpretability of the network by deleting redundant feature nodes and augmenting nodes. The method includes:

[0079] Based on the dynamic features related to the extracted key performance indicator (KPI) data, an additional feature was added to the conventional BLS loss function. The term is used, and the objective function of SBLS is transformed into an unconstrained optimization problem using augmented Lagrange multipliers:

[0080]

[0081] In the formula, It is the output matrix of the fault type. The actual input calculated from the dynamic feature F related to KPIs. These are the weights of SBLS. and These are the ridge parameter and the Lasso parameter, respectively. , , These are intermediate variables for the objective function. It is a square formation. For Lagrange multipliers, It is a positive parameter.

[0082] In one possible implementation, S3 further includes:

[0083] Use block coordinate descent (BCD) to solve for intermediate variables and parameters:

[0084]

[0085] in, The calculation is as follows:

[0086]

[0087] In the formula, , , .

[0088] In one possible implementation, S3 further includes:

[0089] Based on the dynamic characteristics of key performance indicator (KPI) data, and the intermediate variables and parameters used in the solution process, fault diagnosis results are obtained:

[0090] Fault diagnosis results The output of SBLS is calculated as follows:

[0091]

[0092] In the formula, The optimized weight matrix is ​​shown below, where A represents the actual input matrix calculated from the dynamic features F related to the key performance indicator (KPI) data; the KPI-related fault diagnosis structure diagram based on SMGU-SBLS is shown below. Figure 3 As shown.

[0093] S4. Based on the trained Supervised Minimum Gated Unit (SMGU) and Sparse Width Learning System (SBLS), obtain the KPI-related fault diagnosis structure based on SMGU-SBLS.

[0094] S5. Obtain the fault data to be detected, input the fault data to be detected into the KPI-related fault diagnosis structure based on SMGU-SBLS, and obtain the fault diagnosis result.

[0095] In one feasible implementation, in step S5, the fault data to be detected is acquired, and the fault data is input into the SMGU-SBLS-based KPI-related fault diagnosis structure to obtain the fault diagnosis result, including:

[0096] Acquire the fault data to be detected, and the fault type of the fault data;

[0097] Calculate the input matrix and pseudo-inverse matrix for the fault data to be detected;

[0098] Based on the input matrix and pseudo-inverse matrix of the fault data to be detected, the weight matrix of the fault data to be detected is calculated.

[0099] The fault diagnosis results are calculated based on the input matrix and the weight matrix.

[0100] In one feasible implementation, SMGU-SBLS does not require a complex retraining process when new fault data is added, and it has good scalability for multi-mode manufacturing processes.

[0101] In one feasible implementation, new fault data Include A new type of fault, input matrix Its pseudo-inverse matrix is ​​calculated as follows:

[0102]

[0103] In the formula, The augmented matrix of the input matrix. , ,if =0, ,otherwise ;

[0104] The new weight matrix is ​​calculated as follows:

[0105]

[0106] In the formula, It is an augmented matrix of new faults. The extended structure diagram of SMGU-SBLS is shown below, where the matrix is ​​zero. Figure 4 As shown. Figure 5 As shown in the figure, the fault diagnosis accuracy based on SMGU-SBLS is higher than that of the other four methods under different sample sizes.

[0107] In this embodiment of the invention, the KPI-related fault diagnosis method for multi-mode manufacturing processes provided by the present invention will offer new ideas and approaches for fault diagnosis of hot-rolled strip steel based on deep neural networks. It makes up for the complex fault diagnosis problems of multi-mode manufacturing processes, which are difficult to solve by existing fault diagnosis methods due to their poor versatility and low diagnostic accuracy. It aims to explore practical and effective solutions to the key challenges in fault diagnosis of modern industrial processes and has important practical application and promotion value.

[0108] Compared to existing fault diagnosis methods, the fault diagnosis accuracy has been improved by 3% to 5%.

[0109] SMGU-based KPI-related dynamic feature extraction: MGU significantly reduces training complexity and improves training speed by simplifying the gate structure. However, unsupervised MGU mainly learns dynamic features from process variables, which may contain information unrelated to KPIs. Supervised SMGU can fully extract KPI-related dynamic characteristics.

[0110] Fine-grained fault diagnosis based on SBLS: The large number of nodes in BLS increases the complexity of the network. Sparse BLS simplifies the structure of traditional BLS by removing redundant feature nodes and enhancing nodes, effectively avoiding overfitting of the model and enhancing the interpretability of the network.

[0111] Figure 6This is a block diagram of a KPI-related fault diagnosis apparatus 300 for a multi-mode manufacturing process, illustrated according to an exemplary embodiment. The apparatus 300 is used for a KPI-related fault diagnosis method for a multi-mode manufacturing process. (Refer to...) Figure 6 The device includes a data acquisition module 310, a feature extraction module 320, an SBLS construction module 330, an SMGU-SBLS structure construction module 340, and a fault diagnosis module 350. Among them:

[0112] Data acquisition module 310 is used to acquire process variables and key performance indicators (KPIs) data retained during the hot rolling of strip steel and to establish a historical sample database.

[0113] Feature extraction module 320 is used to construct a supervised minimum gated unit (SMGU), train the supervised minimum gated unit SMGU based on a historical sample database, and extract dynamic features related to key performance indicator (KPI) data using the supervised minimum gated unit SMGU.

[0114] SBLS building module 330 is used to build the Sparse Width Learning System SBLS. It inputs the dynamic features related to the extracted key performance indicator (KPI) data into the Sparse Width Learning System SBLS for fine-grained fault diagnosis training.

[0115] SMGU-SBLS structure building module 340 is used to obtain a KPI-related fault diagnosis structure based on SMGU-SBLS based on the trained supervised minimum gate unit SMGU and sparse width learning system SBLS.

[0116] The fault diagnosis module 350 is used to acquire the fault data to be detected, input the fault data to be detected into the KPI-related fault diagnosis structure based on SMGU-SBLS, and obtain the fault diagnosis result.

[0117] Optionally, process variables include: various data sets generated during the hot rolling process of strip steel when producing steel plates of different steel grades and thicknesses;

[0118] Key performance indicators (KPIs) include the width, crown, thickness, and flatness of the strip exiting the finishing mill.

[0119] Optionally, the feature extraction module 320 is used to construct a supervised minimum gated unit (SMGU);

[0120] Based on key performance indicator (KPI) data and process variables, update rules for supervised minimum gated unit (SMGU) are constructed to obtain the network parameters of the supervised minimum gated unit (SMGU).

[0121] The output of the supervised minimum gating unit (SMGU) is used as a dynamic feature related to the key performance indicator (KPI) data.

[0122] Optionally, SBLS building block 330 is used to build the sparse width learning system SBLS;

[0123] Based on the dynamic characteristics of the extracted key performance indicator (KPI) data, augmented Lagrange multipliers are used to perform unconstrained optimization of the objective function of SBLS. The unconstrained optimization includes solving for the intermediate variables and parameters of the objective function using block coordinate descent (BCD).

[0124] Based on the dynamic characteristics related to key performance indicator (KPI) data, and combined with intermediate variables and parameters of the objective function, fault diagnosis results are obtained.

[0125] Optionally, the fault diagnosis results include:

[0126] Fault diagnosis results The output of SBLS is calculated as follows:

[0127]

[0128] in, Here is the weight matrix of SBLS. The weights represent the final weights obtained after several iterations, and A represents the actual input matrix calculated from the dynamic features F related to the key performance indicator (KPI) data.

[0129] Optionally, the fault diagnosis module 350 is used to acquire the fault data to be detected and the fault type of the fault data to be detected;

[0130] The SMGU-SBLS-based KPI-related fault diagnosis structure is used to extract features from the fault data to be detected, and to calculate the input matrix and weight matrix.

[0131] Fault diagnosis results are obtained based on the input matrix and weight matrix.

[0132] In this invention, the KPI-related fault diagnosis device for multi-mode manufacturing processes provides a new approach and method for fault diagnosis of hot-rolled strip steel based on deep neural networks. It overcomes the challenges of existing fault diagnosis methods, which suffer from poor versatility and low diagnostic accuracy in multi-mode manufacturing processes. The device aims to explore practical and effective solutions to key challenges in fault diagnosis of modern industrial processes, and has significant practical application and promotion value. Compared to existing fault diagnosis methods, the fault diagnosis accuracy is improved by 3% to 5%.

[0133] Figure 7This is a schematic diagram of the structure of a KPI-related fault diagnosis device for multi-mode manufacturing processes provided in an embodiment of the present invention, as shown below. Figure 7 As shown, KPI-related fault diagnosis equipment for multi-mode manufacturing processes may include the above-mentioned... Figure 6 The illustrated KPI-related fault diagnosis device for multi-mode manufacturing processes. Optionally, the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes may include a first processor 2001.

[0134] Optionally, the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes may also include a memory 2002 and a transceiver 2003.

[0135] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0136] The following is combined Figure 7 A detailed description of each component of the KPI-related fault diagnosis equipment 410 for multi-mode manufacturing processes is provided below:

[0137] The first processor 2001 is the control center of the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0138] Optionally, the first processor 2001 can perform various functions of the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0139] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 7 CPU0 and CPU1 are shown in the diagram.

[0140] In a specific implementation, as one example, the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes may also include multiple processors, for example... Figure 7 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0141] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0142] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently, and may be accessed via the interface circuit of the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes. Figure 7 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0143] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0144] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 7 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0145] Optionally, the transceiver 2003 can be integrated with the first processor 2001 or exist independently, and can be connected via the interface circuit of the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes. Figure 7(Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0146] It should be noted that, Figure 7 The structure of the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes shown in the figure does not constitute a limitation on the router. The actual knowledge structure identification device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0147] Furthermore, the technical effects of the KPI-related fault diagnosis device 410 for multi-mode manufacturing processes can be referred to the technical effects of the KPI-related fault diagnosis method for multi-mode manufacturing processes described in the above method embodiments, and will not be repeated here.

[0148] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0149] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0150] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0151] It should be understood 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. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0152] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0153] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0154] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0155] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0156] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0157] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A KPI-related fault diagnosis method for multi-mode manufacturing processes, characterized in that, The method includes: S1. Obtain process variables and key performance indicators (KPIs) data retained during the hot rolling process of strip steel, and establish a historical sample database; S2. Construct a supervised minimum gate unit (SMGU), train the SMGU based on the historical sample database, and use the SMGU to extract dynamic features related to the key performance indicator (KPI) data. S3. Construct a Sparse Width Learning System (SBLS) and input the dynamic features related to the extracted Key Performance Indicator (KPI) data into the Sparse Width Learning System (SBLS) for fine-grained fault diagnosis training. S4. Based on the trained Supervised Minimum Gated Unit (SMGU) and Sparse Width Learning System (SBLS), obtain the KPI-related fault diagnosis structure based on SMGU-SBLS. S5. Obtain the fault data to be detected, input the fault data to be detected into the SMGU-SBLS-based KPI-related fault diagnosis structure, and obtain the fault diagnosis result.

2. The KPI-related fault diagnosis method for multi-mode manufacturing processes according to claim 1, characterized in that, In S1, the process variables include: various data sets generated during the hot rolling process of strip steel when producing steel plates of different steel grades and different thicknesses; Key performance indicators (KPIs) include the width, crown, thickness, and flatness of the strip exiting the finishing mill.

3. The KPI-related fault diagnosis method for multi-mode manufacturing processes according to claim 2, characterized in that, In step S2, a supervised minimum gating unit (SMGU) is constructed and trained based on the historical sample database. The SMGU is then used to extract dynamic features related to the key performance indicator (KPI) data, including: Construct a supervised minimum gated unit (SMGU); Based on the key performance indicator (KPI) data and process variables, update rules for the supervised minimum gated unit (SMGU) are constructed, and the network parameters of the supervised minimum gated unit (SMGU) are obtained. The output of the supervised minimum gating unit (SMGU) is used as a dynamic feature related to the key performance indicator (KPI) data.

4. The KPI-related fault diagnosis method for multi-mode manufacturing processes according to claim 3, characterized in that, S3 involves constructing a Sparse Width Learning System (SBLS), inputting the extracted dynamic features related to the key performance indicator (KPI) data into the SBLS for fine-grained fault diagnosis training, including: Constructing a Sparse Width Learning System (SBLS); Based on the dynamic features related to the extracted key performance indicator (KPI) data, augmented Lagrange multipliers are used to perform unconstrained optimization of the objective function of SBLS; wherein, the unconstrained optimization includes: using block coordinate descent (BCD) to solve for the intermediate variables and parameters of the objective function; Based on the dynamic characteristics related to the key performance indicator (KPI) data, and combined with the intermediate variables and parameters of the objective function, the fault diagnosis results are obtained.

5. The KPI-related fault diagnosis method for multi-mode manufacturing processes according to claim 4, characterized in that, The fault diagnosis results include: Fault diagnosis results The output of SBLS is calculated as follows: ； in, Here is the weight matrix of SBLS. The weights represent the final weights obtained after several iterations, and A represents the actual input matrix calculated from the dynamic features F related to the key performance indicator (KPI) data.

6. The KPI-related fault diagnosis method for multi-mode manufacturing processes according to claim 5, characterized in that, In step S5, the fault data to be detected is acquired, and the fault data is input into the SMGU-SBLS-based KPI-related fault diagnosis structure to obtain the fault diagnosis result, including: Acquire the fault data to be detected, and the fault type of the fault data; The SMGU-SBLS-based KPI-related fault diagnosis structure is used to extract features from the fault data to be detected, and to calculate the input matrix and weight matrix. Fault diagnosis results are obtained based on the input matrix and weight matrix.

7. A KPI-related fault diagnosis device for multi-mode manufacturing processes, wherein the KPI-related fault diagnosis device for multi-mode manufacturing processes is used to implement the KPI-related fault diagnosis method for multi-mode manufacturing processes as described in any one of claims 1-6, characterized in that, The device includes: The data acquisition module is used to acquire process variables and key performance indicators (KPIs) data retained during the hot rolling of strip steel and to establish a historical sample database. The feature extraction module is used to construct a supervised minimum gate unit (SMGU), train the SMGU based on the historical sample database, and extract dynamic features related to key performance indicator (KPI) data using the SMGU. The SBLS building module is used to build the Sparse Width Learning System (SBLS). The dynamic features related to the extracted key performance indicator (KPI) data are input into the Sparse Width Learning System (SBLS) for fine-grained fault diagnosis training. The SMGU-SBLS structure building module is used to obtain a KPI-related fault diagnosis structure based on SMGU-SBLS based on the trained supervised minimum gate unit SMGU and sparse width learning system SBLS. The fault diagnosis module is used to acquire the fault data to be detected, input the fault data to be detected into the SMGU-SBLS-based KPI-related fault diagnosis structure, and obtain the fault diagnosis result.

8. The KPI-related fault diagnosis device for multi-mode manufacturing processes according to claim 7, characterized in that, The process variables include: various data sets generated during the hot rolling process of strip steel when producing steel plates of different steel grades and thicknesses; Key performance indicators (KPIs) include the width, crown, thickness, and flatness of the strip exiting the finishing mill.

9. A KPI-related fault diagnosis device for multi-mode manufacturing processes, characterized in that, The KPI-related fault diagnosis equipment for multi-mode manufacturing processes includes: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 1 to 6.

Images