Process quality diagnosis method, system and equipment based on industrial internet of things, and medium

Abstract

The invention discloses a process quality diagnosis method, system and device based on the industrial Internet of Things and a medium, and relates to the technical field of the industrial Internet of Things, and the method comprises the steps: obtaining the process parameter data of each sub-process according to a first preset frequency, and obtaining the workpiece parameter data of a workpiece at each process node according to a second preset frequency; determining a process parameter characteristic matrix according to the process parameter data; determining a single node feature matrix according to the workpiece parameter data, and determining an adjacent node feature matrix according to the workpiece parameter data; and based on a process quality diagnosis model, generating a process quality diagnosis result according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector. The method has the effect of improving the accuracy of procedure quality diagnosis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial Internet of Things, and in particular to a process quality diagnosis method, system, device and medium based on industrial Internet of Things. BACKGROUND

[0002] At present, the process quality diagnosis technology mainly includes statistical process control (SPC), visual detection, sensor monitoring and automatic detection equipment. These technologies can identify abnormal situations in production to a certain extent by monitoring process parameters or product states in real time, for example, collecting temperature, pressure and other data using sensors, identifying surface defects in combination with visual detection, and identifying defects using automatic equipment. These methods are widely used in industrial production and can improve detection efficiency and automation level.

[0003] However, although these technologies play an important role in practical application, the accuracy of current process quality diagnosis still has deficiencies, because traditional detection means are difficult to fully cope with various complex defect types and variable production environments, and are prone to misjudgment or omission, thereby affecting the reliability of diagnosis. Therefore, the current process quality diagnosis technology still needs to be improved in terms of accuracy, and the detection means needs to be further optimized to further improve the reliability and accuracy of its diagnosis. SUMMARY

[0004] In order to improve the accuracy of process quality diagnosis, the present application provides a process quality diagnosis method, system, device and medium based on industrial Internet of Things.

[0005] In a first aspect, the present application provides a process quality diagnosis method based on industrial Internet of Things, which adopts the following technical solution: The process quality diagnosis method based on industrial Internet of Things is applied to an industrial Internet of Things system, the industrial Internet of Things system includes a management platform, a sensor network platform and an object platform which are sequentially connected in communication, the method is executed by the management platform, and includes: According to a first preset frequency, process parameter data of each sub-process is acquired, and according to a second preset frequency, workpiece parameter data of a workpiece at each process node is acquired; According to the process parameter data, a process parameter feature matrix is determined, wherein the process parameter feature matrix is used to represent process parameter change characteristics of a single sub-process; According to the workpiece parameter data, a single node feature matrix is determined, and according to the workpiece parameter data, an adjacent node feature matrix is determined, wherein the single node feature matrix is used to represent workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is used to represent workpiece parameter difference characteristics of the workpiece before and after the sub-process; generate a process quality diagnosis result based on the process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector, wherein the process quality diagnosis result is used to represent process stability, workpiece quality stability and correlation between the process and the workpiece quality corresponding to each sub-process.

[0006] By adopting the technical solution, the process parameter data of each sub-process is obtained according to the first preset frequency, and the workpiece parameter data of the workpiece at each process node is obtained according to the second preset frequency, then the process parameter feature matrix is determined according to the process parameter data, wherein the process parameter feature matrix is used to represent the process parameter change feature of a single sub-process, then the single node feature matrix is determined according to the workpiece parameter data, and the adjacent node feature matrix is determined according to the workpiece parameter data, wherein the single node feature matrix is used to represent the workpiece parameter fluctuation feature of a single process node, and the adjacent node feature matrix is used to represent the workpiece parameter difference feature of the workpiece before and after the sub-process, finally, the process quality diagnosis result is generated based on the process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector, wherein the process quality diagnosis result is used to represent the process stability, the workpiece quality stability and the correlation between the process and the workpiece quality corresponding to each sub-process. Through the above method, multi-frequency and multi-level data acquisition and analysis are realized, which helps to locate the problems in the production process, thereby improving the accuracy and reliability of the process quality diagnosis.

[0007] Optionally, the step of determining the process parameter feature matrix according to the process parameter data comprises: for each sub-process, determining a process parameter matrix corresponding to the sub-process according to the process parameter data, wherein the rows of the process parameter matrix represent each process parameter of the sub-process at different collection points; performing row difference processing on the process parameter matrix to obtain a process parameter fluctuation matrix corresponding to the sub-process; performing matrix normalization processing on the process parameter fluctuation matrix to obtain a first normalized matrix; obtaining a first importance coefficient corresponding to each process parameter in the sub-process, and performing matrix transformation on the first normalized matrix according to the importance coefficient to obtain a process parameter feature matrix corresponding to the sub-process.

[0008] By adopting the technical scheme, in order to determine the process parameter feature matrix, for each sub-process, the process parameter matrix corresponding to the sub-process is determined according to the process parameter data, wherein the rows of the process parameter matrix represent various process parameters of the sub-process at different collection points, then the process parameter matrix is subjected to row difference processing to obtain the process parameter fluctuation matrix corresponding to the sub-process, then the process parameter fluctuation matrix is subjected to matrix normalization processing to obtain the first normalized matrix, then the first importance coefficient corresponding to each process parameter in the sub-process is obtained, and the first normalized matrix is subjected to matrix transformation according to the importance coefficient to obtain the process parameter feature matrix corresponding to the sub-process.

[0009] Optionally, the step of determining the single-node feature matrix according to the workpiece parameter data comprises: for each process node, the workpiece parameter matrix corresponding to the process node is determined according to the process parameter data, wherein the rows of the workpiece parameter matrix represent various workpiece parameters of different workpieces at the process node; the workpiece parameter matrix is subjected to row difference processing to obtain the workpiece parameter fluctuation matrix corresponding to the process node; the single-node intermediate matrix is subjected to matrix normalization processing to obtain the second normalized matrix; the second importance coefficient corresponding to each process parameter in the sub-process is obtained, and the second normalized matrix is subjected to matrix transformation according to the second importance coefficient to obtain the single-node feature matrix corresponding to the process node.

[0010] By adopting the technical scheme, in order to determine the single-node feature matrix, for each process node, the workpiece parameter matrix corresponding to the process node is determined according to the process parameter data, wherein the rows of the workpiece parameter matrix represent various workpiece parameters of different workpieces at the process node, then the workpiece parameter matrix is subjected to row difference processing to obtain the workpiece parameter fluctuation matrix corresponding to the process node, then the single-node intermediate matrix is subjected to matrix normalization processing to obtain the second normalized matrix, and finally the second importance coefficient corresponding to each process parameter in the sub-process is obtained, and the second normalized matrix is subjected to matrix transformation according to the second importance coefficient to obtain the single-node feature matrix corresponding to the process node.

[0011] Optionally, the step of determining the adjacent-node feature matrix according to the workpiece parameter data comprises: for each adjacent process node in each sub-process, the adjacent-node intermediate matrix is obtained by performing matrix subtraction operation on the workpiece parameter matrices corresponding to the adjacent process nodes, wherein the adjacent-node intermediate matrix is used to represent the workpiece parameter difference condition of the workpiece before and after the sub-process; The matrix normalization processing is performed on the intermediate matrix of the adjacent nodes to obtain a third normalized matrix, and the third normalized matrix is transformed according to the second importance coefficient to obtain the adjacent node feature matrix corresponding to the adjacent process nodes.

[0012] By adopting the technical solution, in order to determine the adjacent node feature matrix, for the adjacent process nodes in each sub-process, the matrix subtraction operation is performed according to the workpiece parameter matrices corresponding to the adjacent process nodes respectively to obtain the intermediate matrix of the adjacent nodes, wherein the intermediate matrix of the adjacent nodes is used to represent the workpiece parameter difference condition before and after the sub-process, then the matrix normalization processing is performed on the intermediate matrix of the adjacent nodes to obtain the third normalized matrix, and the third normalized matrix is transformed according to the second importance coefficient to obtain the adjacent node feature matrix corresponding to the adjacent process nodes.

[0013] Optionally, the process quality diagnosis model comprises an input layer, a hidden layer and an output layer, and the step of generating the process quality diagnosis result based on the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector comprises: generating an input feature vector through the input layer according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector; generating a feature extraction vector through the hidden layer according to the input feature vector; generating an output matrix through the output layer according to the feature extraction vector, wherein the dimension of the output matrix is m x 3, m is consistent with the number of the sub-processes, and each column of the output matrix is used to represent a process stability index, a workpiece quality stability index and a process-quality coupling coefficient respectively; generating a network stability analysis result based on the output matrix.

[0014] By adopting the technical solution, in order to generate the process quality diagnosis result, an input feature vector is generated through the input layer according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector, then a feature extraction vector is generated through the hidden layer according to the input feature vector, then an output matrix is generated through the output layer according to the feature extraction vector, wherein the dimension of the output matrix is m x 3, m is consistent with the number of the sub-processes, each column of the output matrix is used to represent a process stability index, a workpiece quality stability index and a process-quality coupling coefficient respectively, and finally a network stability analysis result is generated based on the output matrix.

[0015] Optionally, the hidden layers include a first hidden layer, a second hidden layer and a third hidden layer, a neuron ratio between the first hidden layer, the second hidden layer and the third hidden layer is 8:4:1, and the step of generating a feature extraction vector according to the input feature vector through the hidden layers includes: generating a first intermediate feature vector based on the first hidden layer and according to an input feature vector; generating a second intermediate feature vector based on the second hidden layer and according to the first intermediate feature vector; generating a feature extraction vector based on the third hidden layer and according to the second intermediate feature vector.

[0016] By adopting the above technical solution, in order to generate a feature extraction vector, a first intermediate feature vector is generated based on a first hidden layer and according to an input feature vector, then a second intermediate feature vector is generated based on a second hidden layer and according to the first intermediate feature vector, and finally a feature extraction vector is generated based on a third hidden layer and according to the second intermediate feature vector.

[0017] Optionally, before the step of acquiring process parameter data of each sub-process according to a first preset frequency and acquiring workpiece parameter data of a workpiece at each process node according to a second preset frequency, the method further includes: acquiring processing rates of the sub-processes, determining a corresponding greatest common divisor according to the processing rates, and determining a corresponding least common multiple according to the processing rates; taking the greatest common divisor as the second preset frequency and taking the least common multiple as the first preset frequency.

[0018] By adopting the above technical solution, in order to determine a first preset frequency and a second preset frequency, processing rates of the sub-processes are acquired, a corresponding greatest common divisor is determined according to the processing rates, and a corresponding least common multiple is determined according to the processing rates, then the greatest common divisor is taken as the second preset frequency and the least common multiple is taken as the first preset frequency.

[0019] In a second aspect, the application further provides a process quality diagnosis based on an industrial Internet of Things, and adopts the following technical solution: A process quality diagnosis system based on an industrial Internet of Things includes a management platform, a sensing network platform and an object platform which are sequentially connected in communication, and the management platform is configured to have: a data acquisition module for acquiring process parameter data of each sub-process according to a first preset frequency and acquiring workpiece parameter data of a workpiece at each process node according to a second preset frequency; a parameter feature generation module configured to determine a process parameter feature matrix according to the process parameter data, wherein the process parameter feature matrix is used to represent process parameter variation features of the single sub-process; a node feature generation module configured to determine a single node feature matrix according to the workpiece parameter data and to determine an adjacent node feature matrix according to the workpiece parameter data, wherein the single node feature matrix is used to represent workpiece parameter fluctuation features of a single process node, and the adjacent node feature matrix is used to represent workpiece parameter difference features of the workpiece before and after the sub-process; a process quality diagnosis module configured to generate a process quality diagnosis result based on a process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector, wherein the process quality diagnosis result is used to represent process stability, workpiece quality stability and correlation between the process and the workpiece quality corresponding to each sub-process.

[0020] In a third aspect, the present application further provides a computer device, which adopts the technical scheme as follows: A computer device, comprising a memory and a processor, wherein the memory stores a computer program capable of running on the processor, and the processor implements the method in the first aspect when executing the computer program.

[0021] In a fourth aspect, the present application further provides a computer readable storage medium, which adopts the technical scheme as follows: A computer readable storage medium, which stores a computer program capable of being loaded and executed by a processor to implement the method in the first aspect.

[0022] In summary, the present application at least includes the following beneficial technical effects: first, according to the first preset frequency, the process parameter data of each process is obtained, and according to the second preset frequency, the workpiece parameter data of the workpiece at each process node is obtained, then the process parameter feature matrix is determined according to the process parameter data, wherein the process parameter feature matrix is used to represent the process parameter change characteristics of a single process, then the single node feature matrix is determined according to the workpiece parameter data, and the adjacent node feature matrix is determined according to the workpiece parameter data, wherein the single node feature matrix is used to represent the workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is used to represent the workpiece parameter difference characteristics of the workpiece before and after the process, finally, based on the process quality diagnosis model, and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector, the process quality diagnosis result is generated, wherein the process quality diagnosis result is used to represent the process stability, the workpiece quality stability and the correlation between the process and the workpiece quality corresponding to each process; through the above method, multi-frequency and multi-level data acquisition and analysis are realized, which helps to locate the problems in the production process, thereby improving the accuracy and reliability of the process quality diagnosis. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a whole flow schematic diagram of the embodiment of the present application.

[0024] Figure 2 is a structure schematic diagram of one of the application scenarios of the system of the embodiment of the present application.

[0025] Figure 3 is a structure schematic diagram of another application scenario of the system of the embodiment of the present application.

[0026] Figure 4 is a structure block diagram of the computer device of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application. Figures 1-4

[0028] The embodiment of the present application discloses a process quality diagnosis method based on industrial Internet of Things.

[0029] With reference to Figure 1 , the process quality diagnosis method based on industrial Internet of Things is applied to an industrial Internet of Things system, the industrial Internet of Things system includes a management platform, a sensing network platform and an object platform which are sequentially connected in communication, the method is executed by the management platform, and includes the following steps. ​Step S11, obtaining process parameter data of each sub-process according to a first preset frequency, and obtaining workpiece parameter data of the workpiece at each process node according to a second preset frequency.

[0030] It should be noted that in step S11, the sub-process is a subdivided operation of the process, and the process node is located between the sub-processes. The process node is an important dividing point or detection point in time or space, and connects each sub-process.

[0031] It should be further noted that the first preset frequency is used to obtain the process parameter data of each sub-process, which emphasizes high-frequency monitoring of the parameter changes within the sub-process, reflects the details and dynamic state of each sub-process, and therefore the frequency is relatively high, which facilitates capturing the slight changes or abnormalities in the process within a short period of time. The second preset frequency is used to obtain the workpiece parameter data of the workpiece at each process node, which focuses on the overall state and quality of the workpiece at the key time node or space node, and focuses on the monitoring and control of the overall state of the workpiece. Therefore, the frequency is relatively low. In order to ensure the time consistency and synchronization of the process parameter data and the workpiece parameter data in time, the second preset frequency is an integer multiple of the first preset frequency. Further, in the process of collecting the workpiece parameter data according to the second preset frequency, it must be ensured that the corresponding workpiece has completed the corresponding sub-process and is in a detectable state when collecting data at the process node. This means that the process step corresponding to the collection time point must have been completely completed, and the parameter reading is performed after the workpiece state is stable, so as to ensure that the obtained workpiece parameter truly reflects the overall state of the workpiece after processing, rather than the workpiece in an incomplete or processing state. That is, when the workpiece parameter data is obtained according to the second preset frequency, it should be ensured that there is a workpiece that has been processed at each process node.

[0032] Step S12, determining a process parameter feature matrix according to the process parameter data.

[0033] The process parameter feature matrix is used to represent the process parameter change characteristics of a single sub-process.

[0034] Step S13, determining a single node feature matrix according to the workpiece parameter data, and determining an adjacent node feature matrix according to the workpiece parameter data.

[0035] The single node feature matrix is used to represent the workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is used to represent the workpiece parameter difference characteristics of the workpiece before and after the sub-process.

[0036] Step S14, generating a process quality diagnosis result based on a process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector.

[0037] The process quality diagnosis result is used to represent the process stability corresponding to each sub-process, the workpiece quality stability, and the correlation between the process and the workpiece quality.

[0038] In the above embodiment, the process parameter data of each sub-process is obtained according to the first preset frequency, the workpiece parameter data of the workpiece at each process node is obtained according to the second preset frequency, the process parameter feature matrix is determined according to the process parameter data, the single node feature matrix is determined according to the workpiece parameter data, and the adjacent node feature matrix is determined according to the workpiece parameter data, the process parameter feature matrix is used to represent the process parameter change characteristics of a single sub-process, the single node feature matrix is used to represent the workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is used to represent the workpiece parameter difference characteristics of the workpiece before and after the sub-process, the process quality diagnosis result is generated based on the process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix, and the adjacent node feature vector, and the process quality diagnosis result is used to represent the process stability corresponding to each sub-process, the workpiece quality stability, and the correlation between the process and the workpiece quality. Through the above method, multi-frequency and multi-level data acquisition and analysis are realized, which helps to locate the problems in the production process, thereby improving the accuracy and reliability of the process quality diagnosis.

[0039] As a further embodiment of the method, the step of determining the process parameter feature matrix according to the process parameter data comprises: In step S21, for each sub-process, the process parameter matrix corresponding to the sub-process is determined according to the process parameter data.

[0040] The rows of the process parameter matrix represent the process parameters of the sub-process at different collection points.

[0041] In step S22, the process parameter matrix is subjected to row difference processing to obtain the process parameter fluctuation matrix corresponding to the sub-process.

[0042] It should be noted that the row difference processing highlights the change information in time or space by calculating the difference between consecutive rows (or row vectors), reduces the smooth part or common interference that may exist in the data, and enhances the change characteristics of the signal. The specific implementation is usually to calculate the difference between each row in the data matrix and the previous row to obtain a difference matrix. This operation can be performed in time series or spatial dimensions to capture different levels of change characteristics.

[0043] In step S23, the process parameter fluctuation matrix is subjected to matrix normalization processing to obtain a first normalized matrix.

[0044] It should be noted that matrix normalization is a standardization technique aiming to adjust the values in a data matrix to a uniform scale to eliminate the dimensional differences between different features, facilitating subsequent analysis and modeling; in the processing of multi-source, multi-scale data such as process parameters and workpiece parameters, matrix normalization is particularly important.

[0045] In step S24, the first importance coefficients corresponding to the process parameters in the sub-process are obtained, and the first normalized matrix is subjected to matrix transformation according to the importance coefficients, to obtain the process parameter feature matrix corresponding to the sub-process.

[0046] It should be noted that the first importance coefficients corresponding to the process parameters are usually different in different sub-processes, which is mainly related to the process characteristics and key influencing factors of the sub-process; step S24 highlights the key parameters by weighting the normalized matrix with the first importance coefficients of the key parameters, thereby improving the representativeness and diagnostic effect of the feature matrix.

[0047] In the above embodiment, in order to determine the process parameter feature matrix, for each sub-process, a process parameter matrix corresponding to the sub-process is determined according to the process parameter data, wherein the rows of the process parameter matrix represent the process parameters of the sub-process at different collection points, then the process parameter matrix is subjected to row difference processing to obtain a process parameter fluctuation matrix corresponding to the sub-process, then the process parameter fluctuation matrix is subjected to matrix normalization processing to obtain a first normalized matrix, then the first importance coefficients corresponding to the process parameters in the sub-process are obtained, and the first normalized matrix is subjected to matrix transformation according to the importance coefficients, to obtain the process parameter feature matrix corresponding to the sub-process.

[0048] As a further embodiment of the method, the step of determining a single-node feature matrix according to the workpiece parameter data comprises: In step S31, for each process node, a workpiece parameter matrix corresponding to the process node is determined according to the process parameter data.

[0049] The rows of the workpiece parameter matrix represent the workpiece parameters of different workpieces at the process node.

[0050] In step S32, the workpiece parameter matrix is subjected to row difference processing to obtain a workpiece parameter fluctuation matrix corresponding to the process node.

[0051] In step S33, the single-node intermediate matrix is subjected to matrix normalization processing to obtain a second normalized matrix.

[0052] In step S34, the second importance coefficients corresponding to the process parameters in the sub-process are obtained, and the second normalized matrix is subjected to matrix transformation according to the second importance coefficients, to obtain the single-node feature matrix corresponding to the process node.

[0053] In the above embodiment, in order to determine the single-node feature matrix, for each process node, a workpiece parameter matrix corresponding to the process node is determined according to the process parameter data, wherein a row of the workpiece parameter matrix represents each workpiece parameter of different workpieces at the process node, then the workpiece parameter matrix is subjected to row difference processing to obtain a workpiece parameter fluctuation matrix corresponding to the process node, then the single-node intermediate matrix is subjected to matrix normalization processing to obtain a second normalized matrix, finally, the second importance coefficients corresponding to each process parameter in the sub-process are obtained, and the second normalized matrix is subjected to matrix transformation according to the second importance coefficients to obtain the single-node feature matrix corresponding to the process node.

[0054] As a further embodiment of the method, the step of determining the adjacent-node feature matrix according to the workpiece parameter data comprises: In step S41, for each adjacent process node in each sub-process, a matrix subtraction operation is performed on the workpiece parameter matrices corresponding to the adjacent process nodes to obtain an adjacent-node intermediate matrix.

[0055] The adjacent-node intermediate matrix is used to represent the difference in workpiece parameters before and after the sub-process.

[0056] In step S42, the adjacent-node intermediate matrix is subjected to matrix normalization processing to obtain a third normalized matrix, and the third normalized matrix is subjected to matrix transformation according to the second importance coefficients to obtain the adjacent-node feature matrix corresponding to the adjacent process nodes.

[0057] In the above embodiment, in order to determine the adjacent-node feature matrix, for each adjacent process node in each sub-process, a matrix subtraction operation is performed on the workpiece parameter matrices corresponding to the adjacent process nodes to obtain an adjacent-node intermediate matrix, wherein the adjacent-node intermediate matrix is used to represent the difference in workpiece parameters before and after the sub-process, then the adjacent-node intermediate matrix is subjected to matrix normalization processing to obtain a third normalized matrix, and the third normalized matrix is subjected to matrix transformation according to the second importance coefficients to obtain the adjacent-node feature matrix corresponding to the adjacent process nodes.

[0058] As a further embodiment of the method, the process quality diagnosis model comprises an input layer, a hidden layer and an output layer, and based on the process quality diagnosis model, the step of generating a process quality diagnosis result according to the process parameter feature matrix, the single-node feature matrix and the adjacent-node feature vector comprises: In step S51, an input feature vector is generated by the input layer according to the process parameter feature matrix, the single-node feature matrix and the adjacent-node feature vector.

[0059] In step S52, a feature extraction vector is generated by the hidden layer according to the input feature vector.

[0060] Step S53, generating an output matrix according to the feature extraction vector through the output layer.

[0061] wherein the output matrix has a dimension of m x 3, m is consistent with the number of sub-processes, and each column of the output matrix is used to represent a process stability index, a workpiece quality stability index, and a process-quality coupling coefficient.

[0062] Step S54, generating a network stability analysis result based on the output matrix.

[0063] In the above embodiment, in order to generate a process quality diagnosis result, an input feature vector is generated through the input layer according to the process parameter feature matrix, the single node feature matrix, and the adjacent node feature vector, then a feature extraction vector is generated through the hidden layer according to the input feature vector, and then an output matrix is generated through the output layer according to the feature extraction vector, wherein the output matrix has a dimension of m x 3, m is consistent with the number of sub-processes, and each column of the output matrix is used to represent a process stability index, a workpiece quality stability index, and a process-quality coupling coefficient, and finally a network stability analysis result is generated based on the output matrix.

[0064] As a further embodiment of the method, the hidden layer includes a first hidden layer, a second hidden layer, and a third hidden layer, the neuron ratio between the first hidden layer, the second hidden layer, and the third hidden layer is 8:4:1, and the step of generating the feature extraction vector through the hidden layer according to the input feature vector includes: Step S61, generating a first intermediate feature vector based on the first hidden layer and according to the input feature vector.

[0065] Step S62, generating a second intermediate feature vector based on the second hidden layer and according to the first intermediate feature vector.

[0066] Step S63, generating the feature extraction vector based on the third hidden layer and according to the second intermediate feature vector.

[0067] In the above embodiment, in order to generate the feature extraction vector, a first intermediate feature vector is generated based on the first hidden layer and according to the input feature vector, then a second intermediate feature vector is generated based on the second hidden layer and according to the first intermediate feature vector, and finally the feature extraction vector is generated based on the third hidden layer and according to the second intermediate feature vector.

[0068] As a further embodiment of the method, before the step of collecting process parameter data of each sub-process based on a first preset frequency and collecting workpiece parameter data of the workpiece at each process node based on a second preset frequency, it further includes: Step S71, acquiring a processing rate of each sub-process, and determining a corresponding greatest common divisor according to the processing rate and a corresponding least common multiple according to the processing rate.

[0069] Step S72, the greatest common divisor is taken as the second preset frequency, and the least common multiple is taken as the first preset frequency.

[0070] In the above embodiment, in order to determine the first preset frequency and the second preset frequency, the processing rate of each sub-process is obtained, the corresponding greatest common divisor is determined according to the processing rate, the corresponding least common multiple is determined according to the processing rate, then the greatest common divisor is taken as the second preset frequency, and the least common multiple is taken as the first preset frequency.

[0071] The embodiment of the application further discloses a process quality diagnosis system based on an industrial Internet of Things.

[0072] Reference Figure 2 The process quality diagnosis system based on the industrial Internet of Things comprises a management platform, a sensing network platform and an object platform which are sequentially connected in communication, and the management platform is configured to have: a data acquisition module configured to acquire process parameter data of each sub-process according to a first preset frequency and to acquire workpiece parameter data of a workpiece at each process node according to a second preset frequency; a parameter feature generation module configured to determine a process parameter feature matrix according to the process parameter data, wherein the process parameter feature matrix is used to represent process parameter change characteristics of a single sub-process; a node feature generation module configured to determine a single node feature matrix according to the workpiece parameter data and to determine an adjacent node feature matrix according to the workpiece parameter data, wherein the single node feature matrix is used to represent workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is used to represent workpiece parameter difference characteristics of the workpiece before and after a sub-process; a process quality diagnosis module configured to generate a process quality diagnosis result according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector based on a process quality diagnosis model, wherein the process quality diagnosis result is used to represent process stability, workpiece quality stability and correlation between the process and the workpiece quality corresponding to each sub-process.

[0073] The overall framework of another application scenario of the process quality diagnosis system based on the industrial Internet of Things is as shown in Figure 3As shown, the user platform, the service platform, the management platform, the sensing network platform and the object platform can be sequentially interacted to form a five-platform architecture based on the industrial Internet of Things. The management platform is configured with a data acquisition module, a parameter feature generation module, a node feature generation module and a process quality diagnosis module. The service platform includes a service total database, n service sub-platforms and n service sub-databases. Each service sub-platform can communicate with the corresponding service sub-database, and each service sub-database can communicate with the service total database. The sensing network platform includes a sensing total database and n sensing network sub-platforms. Each sensing network sub-platform is provided with a sensing sub-database, and each sensing network sub-platform can communicate with the sensing total database.

[0074] Specifically, in another application scenario described above, the process quality diagnosis system based on the industrial Internet of Things includes a management platform, which is configured to: acquire process parameter data of each sub-process according to a first preset frequency, and acquire workpiece parameter data of a workpiece at each process node according to a second preset frequency; determine a process parameter feature matrix according to the process parameter data, wherein the process parameter feature matrix is used to represent process parameter change characteristics of a single sub-process; determine a single node feature matrix according to the workpiece parameter data, and determine an adjacent node feature matrix according to the workpiece parameter data, wherein the single node feature matrix is used to represent workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is used to represent workpiece parameter difference characteristics of the workpiece before and after the sub-process; and generate a process quality diagnosis result based on a process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector, wherein the process quality diagnosis result is used to represent process stability corresponding to each sub-process, workpiece quality stability and correlation between the process and the workpiece quality.

[0075] Through the interaction between the functional platforms of the three-platform or five-platform based process quality diagnosis system based on the industrial Internet of Things, a perfect closed-loop information operation logic is established to ensure the orderly operation of the sensing information and the control information, and the intelligent management of the equipment is realized.

[0076] The process quality diagnosis system based on the industrial Internet of Things can implement any one of the process quality diagnosis methods based on the industrial Internet of Things, and the specific working process of the process quality diagnosis system based on the industrial Internet of Things can refer to the corresponding process in the process quality diagnosis method based on the industrial Internet of Things.

[0077] The embodiment of the application further discloses a computer device.

[0078] Reference Figure 4A computer device includes a memory and a processor, the memory has a computer program capable of running on the processor, and the processor implements any one of the above process quality diagnosis methods based on industrial internet of things when executing the computer program.

[0079] The embodiment of the present application further discloses a computer readable storage medium.

[0080] A computer readable storage medium stores a computer program capable of being loaded and executed by a processor to execute any one of the above process quality diagnosis methods based on industrial internet of things.

[0081] The computer readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, device or apparatus; the program code contained in the computer readable medium can be transmitted by any appropriate medium, including but not limited to wireless, wire, optical cable, RF, etc., or any appropriate combination of the above.

[0082] The above are preferred embodiments of the present application, and are not intended to limit the protection scope of the present application, any feature disclosed in the specification (including the abstract and drawings) can be replaced by other equivalent or similar features unless specifically described, that is, each feature is only an example of a series of equivalent or similar features.

Claims

1. A process quality diagnosis method based on the Industrial Internet of Things, characterized in that, Applied to an industrial Internet of Things (IIoT) system, the IIoT system includes a management platform, a sensor network platform, and an object platform that are sequentially and communicatively connected. The method is executed by the management platform and includes: The process parameter data of each sub-process is obtained according to the first preset frequency, and the workpiece parameter data of the workpiece at each process node is obtained according to the second preset frequency. A process parameter feature matrix is ​​determined based on the process parameter data, wherein the process parameter feature matrix is ​​used to represent the process parameter variation characteristics of a single sub-process; A single node feature matrix is ​​determined based on the workpiece parameter data, and an adjacent node feature matrix is ​​determined based on the workpiece parameter data. The single node feature matrix is ​​used to represent the workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is ​​used to represent the workpiece parameter difference characteristics before and after the sub-process. Based on the process quality diagnosis model, and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vector, the process quality diagnosis result is generated. The process quality diagnosis result is used to represent the process stability, workpiece quality stability and the correlation between the process and the workpiece quality corresponding to each sub-process.

2. The process quality diagnosis method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of determining the process parameter feature matrix based on the process parameter data includes: For each of the sub-processes, a process parameter matrix corresponding to the sub-process is determined based on the process parameter data, wherein the rows of the process parameter matrix represent the various process parameters of the sub-process at different collection points; Perform row difference processing on the process parameter matrix to obtain the process parameter fluctuation matrix corresponding to the sub-process; The process parameter fluctuation matrix is ​​normalized to obtain the first normalized matrix; Obtain the first importance coefficient corresponding to each process parameter in the sub-process, and perform matrix transformation on the first normalized matrix according to the importance coefficient to obtain the process parameter feature matrix corresponding to the sub-process.

3. The process quality diagnosis method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of determining the single-node feature matrix based on the workpiece parameter data includes: For each process node, a workpiece parameter matrix corresponding to the process node is determined based on the process parameter data, wherein the rows of the workpiece parameter matrix represent the various workpiece parameters of different workpieces at the process node; Perform row difference processing on the workpiece parameter matrix to obtain the workpiece parameter fluctuation matrix corresponding to the process node; The intermediate matrix of the single node is normalized to obtain the second normalized matrix; Obtain the second importance coefficients corresponding to each process parameter in the sub-process, and perform matrix transformation on the second normalized matrix based on the second importance coefficients to obtain the single node feature matrix corresponding to the process node.

4. The process quality diagnosis method based on the Industrial Internet of Things according to claim 3, characterized in that, The step of determining the feature matrix of adjacent nodes based on the workpiece parameter data includes: For each of the adjacent process nodes in the sub-process, a matrix subtraction operation is performed based on the workpiece parameter matrix corresponding to each of the adjacent process nodes to obtain the intermediate matrix of the adjacent nodes, wherein the intermediate matrix of the adjacent nodes is used to represent the difference in workpiece parameters before and after the sub-process. The intermediate matrix of the adjacent nodes is normalized to obtain a third normalized matrix. The third normalized matrix is ​​then transformed according to the second importance coefficient to obtain the adjacent node feature matrix corresponding to the adjacent process nodes.

5. The process quality diagnosis method based on the Industrial Internet of Things according to claim 1, characterized in that, The process quality diagnosis model includes an input layer, a hidden layer, and an output layer. The step of generating process quality diagnosis results based on the process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix, and the neighboring node feature vectors includes: Through the input layer, an input feature vector is generated based on the process parameter feature matrix, the single node feature matrix, and the adjacent node feature vectors; Through the hidden layer, a feature extraction vector is generated based on the input feature vector; Through the output layer, an output matrix is ​​generated based on the feature extraction vector, wherein the dimension of the output matrix is ​​m×3, where m is the same as the number of sub-processes, and each column of the output matrix is ​​used to represent the process stability index, the workpiece quality stability index, and the process-quality coupling coefficient, respectively. Based on the output matrix, network stability analysis results are generated.

6. The process quality diagnosis method based on the Industrial Internet of Things according to claim 5, characterized in that, The hidden layer includes a first hidden layer, a second hidden layer, and a third hidden layer, with a neuron ratio of 8:4:1 among the first, second, and third hidden layers. The step of generating a feature extraction vector based on the input feature vector through the hidden layers includes: Based on the first hidden layer, and according to the input feature vector, a first intermediate feature vector is generated; Based on the second hidden layer, and according to the first intermediate feature vector, a second intermediate feature vector is generated; Based on the third hidden layer and according to the second intermediate feature vector, a feature extraction vector is generated.

7. The process quality diagnosis method based on the Industrial Internet of Things according to claim 1, characterized in that, Before the steps of obtaining process parameter data for each sub-process according to a first preset frequency and obtaining workpiece parameter data for each process node according to a second preset frequency, the method further includes: Obtain the processing rate of each sub-process, determine the corresponding greatest common divisor based on the processing rate, and determine the corresponding least common multiple based on the processing rate; The greatest common divisor is used as the second preset frequency, and the least common multiple is used as the first preset frequency.

8. A process quality diagnosis system based on the Industrial Internet of Things, characterized in that, It includes a management platform, a sensor network platform, and an object platform that are connected in sequence. The management platform is configured with: The data acquisition module is used to acquire process parameter data of each sub-process according to a first preset frequency, and to acquire workpiece parameter data of the workpiece at each process node according to a second preset frequency. The parameter feature generation module is used to determine the process parameter feature matrix based on the process parameter data, wherein the process parameter feature matrix is ​​used to represent the process parameter variation characteristics of a single sub-process; The node feature generation module is used to determine a single node feature matrix based on the workpiece parameter data, and to determine adjacent node feature matrices based on the workpiece parameter data. The single node feature matrix is ​​used to represent the workpiece parameter fluctuation characteristics of a single process node, and the adjacent node feature matrix is ​​used to represent the workpiece parameter difference characteristics before and after the sub-process. The process quality diagnosis module is used to generate process quality diagnosis results based on the process quality diagnosis model and according to the process parameter feature matrix, the single node feature matrix and the adjacent node feature vectors. The process quality diagnosis results are used to represent the process stability, workpiece quality stability and correlation between the process and the workpiece quality corresponding to each sub-process.

9. A computer device, characterized in that, The method includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer program is stored that can be loaded by a processor and execute the method of any one of claims 1 to 7.

Images