Quality index alignment method and system between process nodes and electronic equipment

Abstract

The invention relates to the technical field of process indexes, and discloses a quality index alignment method and system between process nodes and electronic equipment. According to the invention, the sampling data between the initial process node and the target process node is calculated through the process duration function to obtain the alternative quality index, and prediction is carried out through the index prediction model according to the initial quality index corresponding to the initial process node to obtain the predicted quality index. The process duration function is updated based on the difference between the alternative quality index and the predicted quality index, so that the target quality index corresponding to the initial quality index is output according to the updated target alignment model, and the difference between the alternative quality index and the predicted quality index is eliminated by updating the process duration function. And aligning the updated index with the model to obtain the aligned target quality index, thereby greatly improving the alignment accuracy of the quality index and accurately reflecting the production condition of the process node.

Description

Technical Field

[0001] The present invention relates to the field of process indicator technology, and in particular to a method, system and electronic equipment for aligning quality indicators between process nodes. Background Art

[0002] In industries such as chemicals, pharmaceuticals, petroleum refining and food processing, the production process from raw material input to product output is complex and sophisticated, and requires multiple continuous or intermittent process nodes, such as mixing, heating, cooling, chemical reactions, material separation and refining and purification. Each of these process nodes carries a specific production task, and uses a series of quality indicators to measure its operating results and product quality. Therefore, ensuring the consistency of quality indicators between nodes is crucial to the quality assurance of the final product. Traditionally, in order to achieve precise control of the production process, the industry generally adopts a static alignment method to manage the quality indicators between each process node. Based on the preset association rules, this method aligns the quality indicators of the next process node with the corresponding indicators of the previous node, so as to monitor and guide the process according to the aligned quality indicators to ensure that production proceeds as expected.

[0003] However, due to the complex interactions between process nodes, the continuous improvement and updating of process data, etc., slight changes in raw materials and process conditions may not be immediately reflected in the instant sampling data, but may gradually accumulate and amplify in the subsequent production stages, and ultimately have a significant impact on product quality. Even for two adjacent process nodes, the correlation between their quality indicators may change due to these dynamic factors.

[0004] Therefore, the quality indicators obtained by relying on static alignment are often less accurate and difficult to accurately reflect the actual production conditions, which in turn leads to deviations in quality control and affects the final quality of the product. Summary of the invention

[0005] In order to provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. The summary is not an extensive review, nor is it intended to identify key / critical components or delineate the scope of protection of these embodiments, but rather serves as a prelude to the detailed description that follows.

[0006] In view of the above-mentioned shortcomings of the prior art, the present application provides a method, system and electronic device for aligning quality indicators between process nodes to improve the alignment accuracy of quality indicators.

[0007] The present application provides a method for aligning quality indicators between process nodes, including: pre-setting an indicator alignment model between an initial process node and a target process node, the indicator alignment model including a process duration function and an indicator prediction model, wherein the process duration function is used to characterize the correspondence between process duration and probability density; obtaining an initial quality indicator corresponding to the initial process node, and performing quality detection in a time period between the initial process node and the target process node to obtain sampled data; predicting using the indicator prediction model according to the initial quality indicator to obtain a predicted quality indicator corresponding to the target process node, and calculating the sampled data according to the process duration function to obtain an alternative quality indicator corresponding to the target process node, so as to update the process duration function according to the difference between the predicted quality indicator and the alternative quality indicator; outputting the alternative quality indicator through the updated indicator alignment model as the target quality indicator aligned with the initial quality indicator.

[0008] In one embodiment of the present application, the process duration function is obtained in the following manner: an expert knowledge base is established in advance for the current process flow, and historical process data corresponding to the current process flow is collected, wherein the current process flow includes the initial process node and the target process node; a process duration library is generated in combination with the expert knowledge base and / or the historical process data, wherein the process duration library stores multiple process scenario data and original duration functions corresponding to each of the process scenario data; current scenario data is obtained, and matching is performed from the process duration library according to the matching results between the current scenario data and each of the process scenario data to obtain a process duration function.

[0009] In one embodiment of the present application, the alternative quality indicator is determined by the following formula: Y = ∑y(t+Δt)×f(Δt); wherein Y is the alternative quality indicator, y(t+Δt) is the sampling data, f(Δt) is the process duration function, t is the initial time point corresponding to the initial process node, and Δt is the process duration between the initial process node and the target process node.

[0010] In one embodiment of the present application, an indicator prediction model is obtained in the following manner: obtaining the raw material quality indicator corresponding to the initial process node, and obtaining the product quality indicator corresponding to the target process node, wherein there is a correlation between the historical process data, the raw material quality indicator and the product quality indicator; establishing an indicator prediction function between the raw material quality indicator and the product quality indicator in combination with the expert knowledge base; using the raw material quality indicator and the historical process data associated with the raw material quality indicator as training samples, and using the product quality indicator associated with the raw material quality indicator as the sample label of the training sample, performing model training on the original prediction model generated based on the indicator prediction function according to the training samples with sample labels, and using the trained original prediction model as the indicator prediction model.

[0011] In one embodiment of the present application, the process duration function is updated according to the difference between the predicted quality index and the alternative quality index, including: if there are multiple process duration functions, then the independent difference index corresponding to each alternative quality index is calculated respectively according to each predicted quality index, and the global difference index is obtained by calculating according to each independent difference index; if the global difference index is greater than or equal to a preset global index threshold, then the process duration functions are classified according to the process scenario to obtain the same-scenario functions corresponding to different process scenarios, wherein the same-scenario functions corresponding to the same process scenario are the same ; Take any process scenario as the target scenario, calculate the predicted quality index and target quality index corresponding to the same target scenario, and obtain the scenario difference index corresponding to the target scenario; if the scenario difference index is greater than or equal to the preset scene index threshold, update the same-scenario functions corresponding to the target scenario; if the update of each process duration function is completed, count the number of updates corresponding to the process duration function, and update the global difference index according to the updated process duration function; if the global difference index is less than the global indicator threshold, or the number of updates is greater than or equal to the preset number threshold, the model update status of the indicator alignment model is determined as update completion.

[0012] In one embodiment of the present application, an independent difference index is calculated in the following manner: one or more difference index types are obtained, wherein the difference index types include one or more of mean square error, root mean square error and determination coefficient; the prediction quality index and the target quality index are calculated according to each of the difference index types, respectively, to obtain independent difference indicators corresponding to each of the difference index types.

[0013] In one embodiment of the present application, each same-scene function corresponding to the target scene is updated, including: generating multiple intermediate functions corresponding to any same-scene function by adjusting function parameters, wherein the interval length corresponding to the process time interval in the intermediate function is smaller than the interval length corresponding to the process time interval in the same-scene function; substituting each of the intermediate functions into each same-scene function corresponding to the target scene, so as to calculate the scene difference index corresponding to the target scene, and obtain the intermediate difference index corresponding to each of the intermediate functions; determining the target function from each of the intermediate functions according to the comparison result between the intermediate difference indexes, and updating the same-scene function according to the target function.

[0014] In one embodiment of the present application, the method further includes: obtaining a current process flow, wherein the current process flow includes a plurality of original process nodes arranged in a preset process order, and an indicator alignment model is set between any two original process nodes; determining the current process flow as a test flow; in response to the test flow, determining the original process node with the first order in the test flow as the test initial node, determining the original process node with the last order in the test flow as the test termination node, determining the indicator alignment model between two adjacent original process nodes in the test flow as the first alignment model, and determining the indicator alignment model between the test initial node and the test termination node as the second alignment model; obtaining a quality indicator sample corresponding to the test initial node; and using each of the first alignment models in turn to obtain the first test indicator sample corresponding to the test termination node. test indicator, and using the second alignment model, obtain the second test indicator corresponding to the quality indicator sample at the test termination node; perform difference calculation based on the first test indicator and the second test indicator to obtain a model normal indicator; if the model normal indicator is less than a preset normal indicator threshold, adjust the test process through a preset adjustment strategy to obtain a new test process, wherein the preset adjustment strategy includes deleting the test initial node from the test process, or deleting the test termination node from the test process; if the model normal indicator is greater than or equal to the normal indicator threshold, determine the abnormal model from the indicator alignment model according to the comparison result between the current process flow and the test process, and update the abnormal model, wherein the abnormal model belongs to the current process flow, and the abnormal model does not belong to the test process.

[0015] The present application provides a quality indicator alignment system between process nodes, including: a setting module, used to set an indicator alignment model between an initial process node and a target process node, the indicator alignment model including a process duration function and an indicator prediction model, wherein the process duration function is used to characterize the correspondence between the process duration and the probability density; an acquisition module, used to acquire an initial quality indicator corresponding to the initial process node, and perform quality detection in a time period between the initial process node and the target process node to obtain sampling data; an update module, used to predict according to the initial quality indicator using the indicator prediction model to obtain a predicted quality indicator corresponding to the target process node, and calculate the sampling data according to the process duration function to obtain an alternative quality indicator corresponding to the target process node, so as to update the process duration function according to the difference between the predicted quality indicator and the alternative quality indicator; an output module, used to output an alternative quality indicator through the updated indicator alignment model as the target quality indicator aligned with the initial quality indicator.

[0016] The present application provides an electronic device, comprising: a processor and a memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device executes the above method.

[0017] Beneficial effects of this application:

[0018] The sampling data between the initial process node and the target process node is calculated through the process duration function to obtain the alternative quality index, and the initial quality index corresponding to the initial process node is predicted through the index prediction model to obtain the predicted quality index, so as to update the process duration function based on the difference between the alternative quality index and the predicted quality index, thereby outputting the target quality index corresponding to the initial quality index according to the updated alignment model. In this way, the quality index of the target process node is assumed according to the alternative quality index obtained from the process duration function and the sampling data, and the predicted quality index obtained according to the index prediction model is used as a reference for the alignment index, thereby eliminating the difference between the alternative quality index and the predicted quality index by updating the process duration function, and obtaining the aligned target quality index by the updated index alignment model, thereby greatly improving the alignment accuracy of the quality index and accurately reflecting the production status of the process node. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flow chart of a method for aligning quality indicators between process nodes in an embodiment of the present application;

[0020] Figure 2is a schematic diagram of the relationship between an initial process node and a target process node in an embodiment of the present application;

[0021] Figure 3 It is a schematic diagram of a process time function in an embodiment of the present application;

[0022] Figure 4 is a schematic diagram of another process time function in an embodiment of the present application;

[0023] Figure 5 It is a schematic diagram of a process time function and its intermediate function in an embodiment of the present application;

[0024] Figure 6 is another flowchart of a method for aligning quality indicators between process nodes in an embodiment of the present application;

[0025] Figure 7 It is a structural schematic diagram of a quality indicator alignment system between process nodes in an embodiment of the present application;

[0026] Figure 8 It is a schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and sub-samples in the embodiments can be combined with each other without conflict.

[0028] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0029] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0030] The terms "first", "second", etc. in the specification and claims of the embodiments of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so that the embodiments of the embodiments of the present disclosure described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions.

[0031] Unless otherwise stated, the term "plurality" means two or more.

[0032] In the embodiment of the present disclosure, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B indicates: A or B.

[0033] The term "and / or" is a description of the association relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or, A and B.

[0034] Combination Figure 1 As shown, the embodiment of the present disclosure provides a method for aligning quality indicators between process nodes, including:

[0035] Step S101, presetting an index alignment model between an initial process node and a target process node;

[0036] Among them, the indicator alignment model includes the process duration function and the indicator prediction model;

[0037] Among them, the process duration function is used to characterize the corresponding relationship between process duration and probability density;

[0038] Step S102, obtaining an initial quality indicator corresponding to an initial process node, and performing quality inspection in a time period between the initial process node and a target process node to obtain sampling data;

[0039] Step S103, predicting using the indicator prediction model according to the initial quality indicator to obtain the predicted quality indicator corresponding to the target process node, and calculating the sampled data according to the process duration function to obtain the alternative quality indicator corresponding to the target process node, so as to update the process duration function according to the difference between the predicted quality indicator and the alternative quality indicator;

[0040] Step S104: Outputting the candidate quality indicator through the updated indicator alignment model as the target quality indicator to which the initial quality indicator is aligned.

[0041] The method for aligning quality indicators between process nodes provided by the embodiment of the present disclosure is adopted. The sampling data between the initial process node and the target process node is calculated through the process duration function to obtain the alternative quality indicator, and the indicator prediction model is used to predict the initial quality indicator corresponding to the initial process node to obtain the predicted quality indicator, so as to update the process duration function based on the difference between the alternative quality indicator and the predicted quality indicator, thereby outputting the target quality indicator corresponding to the initial quality indicator according to the updated alignment model. In this way, the quality indicator of the target process node is assumed according to the alternative quality indicator obtained from the process duration function and the sampling data, and the predicted quality indicator obtained according to the indicator prediction model is used as a reference for the alignment indicator, so as to eliminate the difference between the alternative quality indicator and the predicted quality indicator by updating the process duration function, and the updated indicator alignment model is used to obtain the aligned target quality indicator, thereby greatly improving the alignment accuracy of the quality indicator and accurately reflecting the production status of the process node.

[0042] Optionally, the process duration function is obtained in the following manner: an expert knowledge base is established in advance for the current process flow, and historical process data corresponding to the current process flow is collected, wherein the current process flow includes an initial process node and a target process node; a process duration library is generated by combining the expert knowledge base and the historical process data, wherein the process duration library stores multiple process scenario data and original duration functions corresponding to each process scenario data; the current scenario data is obtained, and according to the matching results between the current scenario data and each process scenario data, matching is performed from the process duration library to obtain the process duration function.

[0043] In this way, the process time function is determined by the expert knowledge base, which ensures that the assumed error of the quality indicator is within a reasonable range, and the indicator prediction model is obtained through the machine learning algorithm, so as to generate an iterative calculation formula based on indicator assumptions and indicator predictions according to the process time function and the indicator prediction model, so as to match the corresponding target quality indicator according to the initial quality indicator, thereby greatly improving the alignment accuracy of the quality indicator and accurately reflecting the production status of the process node.

[0044] Combination Figure 2As shown, an embodiment of the present disclosure provides a schematic diagram of the relationship between an initial process node and a target process node; the initial process node includes placing raw materials into process equipment, wherein an initial quality index corresponding to the initial process node is obtained by sampling and testing the raw materials, and the sampling time point before the raw materials enter the process equipment is used as the initial time point corresponding to the initial process node; the target process node includes outputting products from the process equipment, wherein the time point after the product is output from the process equipment is used as the target time point corresponding to the target process time, and the target quality index corresponding to the target process node is used to characterize the expected quality of the product at the target time point; the process duration is the time difference between the initial time point and the target time point, and the process duration function satisfies a Gaussian distribution with an expected value of μ=4.5 and a standard deviation of σ=0.2; if the initial time point corresponding to the initial process node is t=10, the target time point corresponding to the target process time should be within the Gaussian distribution interval with an expected value of μ=14.5 and a standard deviation of σ=0.2.

[0045] In some embodiments, the historical process data includes raw material formula, equipment status, process parameters, raw material sampling time points, sampling results corresponding to the raw material sampling time points, etc.

[0046] In some embodiments, experts divide the production scenarios according to the on-site production equipment, production processes and other factors to obtain different process scenario data, and set the probability density functions corresponding to the process durations based on the different process scenario data to obtain the original duration functions, wherein the function parameters in the original duration functions can be obtained by relying on expert experience or by analyzing historical process data; the process duration library is shown in Table 1, where if the equipment state is A1, the process parameters are B1 and the other parameters are C1, then the corresponding original duration function is f1(Δt); if the equipment state is A2, the process parameters are B2 and the other parameters are C2, then the corresponding original duration function is f2(Δt); if the equipment state is A1, the process parameters are B2 and the other parameters are C1, then the corresponding original duration function is f3(Δt), wherein other parameters include raw material formula and sampling results, etc.

[0047] Table 1

[0048] Device Status Process parameters Other parameters Original duration function A1 B1 C1 f1(Δt) A2 B2 C2 f2(Δt) A1 B2 C1 <![CDATA[f3(Δt)]]>

[0049] In some embodiments, the original duration function is a probability distribution function, and the probability distribution type of the original duration function includes a Gaussian distribution type, a uniform distribution type, a binomial distribution type, etc.; the original duration function of the Gaussian distribution type is as follows: Figure 3 As shown, the expected value of the original duration function μ = 4.5, the standard deviation of the original duration function σ = 0.2; the original duration function of the uniform distribution type is as follows Figure 4As shown, the process duration in the original duration function is evenly distributed in the interval [4h, 5h].

[0050] In some embodiments, since the original duration function is mapped to the process duration library according to the process scene data, even if the initial time points t are different, since each initial time point has independent current scene data, different time points may correspond to the same process duration function.

[0051] Optionally, the alternative quality indicator is determined by the following formula:

[0052] Y=∑y(t+Δt)×f(Δt) Formula (1)

[0053] In formula (1), Y is the alternative quality indicator, y(t+Δt) is the sampling data, f(Δt) is the process time function, t is the initial time point corresponding to the initial process node, and Δt is the process time between the initial process node and the target process node.

[0054] In some embodiments, the initial time point corresponding to the initial process node is denoted as t, the initial quality index corresponding to the initial process node is denoted as X, the process duration function corresponding to the process duration Δt is denoted as f(Δt), and the target time point is denoted as t+Δt, [t, t+Δt] is the processing time period between the initial process node and the target process node; the sampling data y(t+Δt) is obtained by sampling the processing time period between the initial process node and the target process node; the alternative quality index Y is obtained by calculation based on the process duration function f(Δt) and the sampling data y(t+Δt) as shown in Table 2.

[0055] Table 2

[0056]

[0057] Optionally, the indicator prediction model is obtained in the following manner: obtaining the raw material quality indicator corresponding to the initial process node, and obtaining the product quality indicator corresponding to the target process node, wherein there is a correlation between the historical process data, the raw material quality indicator and the product quality indicator; establishing an indicator prediction function between the raw material quality indicator and the product quality indicator in combination with the expert knowledge base; using the raw material quality indicator and the historical process data associated with the raw material quality indicator as training samples, and using the product quality indicator associated with the raw material quality indicator as the sample label of the training sample, performing model training on the original prediction model generated based on the indicator prediction function according to the training samples with the sample labels, and using the trained original prediction model as the indicator prediction model.

[0058] In some embodiments, an indicator prediction function between raw material quality indicators and product quality indicators is established in combination with an expert knowledge base, wherein the general writing of the indicator prediction function is as follows:

[0059] Y′=G(X) Formula (2)

[0060] In formula (2), Y′ is the product quality index or the predicted quality index, X is the raw material quality index or the initial quality index, and G() is the expression of the index prediction function.

[0061] In some embodiments, the expression of the indicator prediction function is a generalized linear expression, as shown below:

[0062] Y′=∑k i ·x i Formula (3)

[0063] In formula (3), x i is the i-th initial quality index, k i is the adjustment coefficient corresponding to the i-th initial quality index.

[0064] In some embodiments, the expert knowledge base is used to determine whether the relationship between each initial quality indicator and the predicted quality is positive feedback or negative feedback, thereby generating constraints corresponding to each adjustment coefficient. For example, if the initial quality indicator x1 is positive feedback to the predicted quality indicator, then k1>0; if the initial quality indicator x2 is positive feedback to the predicted quality indicator, then k2<0.

[0065] In some embodiments, the original prediction model corresponding to the indicator prediction function is trained and optimized through a machine learning algorithm, wherein the training samples include raw material quality indicators, equipment status, process parameters, etc., and the training labels are product quality indicators.

[0066] In some embodiments, prediction is performed based on the initial quality indicator according to the indicator prediction model, and the obtained predicted quality indicator is shown in Table 3.

[0067] Table 3

[0068]

[0069] Optionally, the process duration function is updated according to the difference between the predicted quality index and the alternative quality index, including: if there are multiple process duration functions, each predicted quality index is calculated with the corresponding alternative quality index to obtain independent difference indicators corresponding to each alternative quality index, and each independent difference index is calculated to obtain a global difference index; if the global difference index is greater than or equal to a preset global index threshold, each process duration function is classified according to the process scenario to obtain same-scenario functions corresponding to different process scenarios, wherein the same-scenario functions corresponding to the same process scenario are the same; any A process scenario is used as a target scenario, and the predicted quality index and the target quality index corresponding to the same target scenario are calculated to obtain a scenario difference index corresponding to the target scenario; if the scenario difference index is greater than or equal to a preset scenario index threshold, the same-scenario functions corresponding to the target scenario are updated; if the process duration functions are updated, the number of updates corresponding to the process duration functions is counted, and the global difference index is updated according to the updated process duration functions; if the global difference index is less than the global index threshold, or the number of updates is greater than or equal to the preset number threshold, the model update status of the indicator alignment model is determined as update completion.

[0070] In some embodiments, it is necessary to calculate the differences between all candidate quality indicators Y and the predicted quality indicator Y'. When MSE is used as the difference indicator, When the MSE for all Y and Y' is less than the set threshold, or the number of iterations reaches the set value, the iteration can be stopped, otherwise the process time function of each category needs to be updated separately.

[0071] In some embodiments, since the same scene functions corresponding to the same process scene are the same, if t0 and t2 have the same process scene, the corresponding process duration function f0 (Δt) and process duration function f2 (Δt) are exactly the same, that is, the process duration function f0 (Δt) and the process duration function f2 (Δt) belong to the same scene function. Assuming that the same scene functions are Gaussian distribution types with μ = 4.5 and σ = 0.2, the calculation formula of the difference index is as follows:

[0072] MSE=(Y′0-Y0) 2 +(Y′2-Y2) 2 Formula (4)

[0073] In formula (4), MSE is the mean square error in the difference index, where if the MSE calculated for t0 and t2 is greater than the preset difference index threshold, the same scene functions f0(Δt) and f2(Δt) are updated synchronously.

[0074] Optionally, the independent difference index is calculated in the following manner: obtaining one or more difference index types, wherein the difference index types include one or more of mean square error, root mean square error and determination coefficient; calculating the prediction quality index and the target quality index according to each difference index type, respectively, to obtain the independent difference index corresponding to each difference index type.

[0075] In some embodiments, MSE (Mean Squared Error), RMSE (Root Mean Square Error), R 2 The indicator alignment model is comprehensively evaluated by using indicators such as Coefficient of Determination (or R-squared) to characterize data differences, so as to determine whether the process duration function should be updated.

[0076] Optionally, each same-scene function corresponding to the target scene is updated, including: generating multiple intermediate functions corresponding to any same-scene function by adjusting function parameters, wherein the interval length corresponding to the process time interval in the intermediate function is smaller than the interval length corresponding to the process time interval in the same-scene function; substituting each of the intermediate functions into each same-scene function corresponding to the target scene, so as to calculate the scene difference index corresponding to the target scene, and obtain the intermediate difference index corresponding to each of the intermediate functions; determining the target function from each of the intermediate functions according to the comparison result between the intermediate difference indexes, and updating the same-scene function according to the target function.

[0077] In some embodiments, if the difference index satisfies the difference index threshold, the target quality index is output through the index alignment model.

[0078] In some embodiments, grouping is performed according to process scenarios; for each group of data, the interval of process duration is narrowed according to different coefficient adjustment strategies, and the best set of probability density functions is selected as the updated probability density function.

[0079] In some embodiments, the process time function f1(Δt) satisfies a Gaussian distribution with an expected value μ=4.5 and a standard deviation σ=0.2; if the difference index is less than the difference index threshold, the standard deviation σ=0.2 is changed to σ=0.1 to narrow the process time interval, and the updated standard deviation is determined according to the expected value μ=4.5, including μ=4.4, μ=4.5, and μ=4.6, to obtain the intermediate functions f1(Δt) and f2(Δt). 1_1 (Δt), intermediate function f 1_2 (Δt), intermediate function f 1_3(Δt), where the intermediate function is as follows Figure 5 As shown; through the intermediate function f 1_1 (Δt) Calculate the sampling data to obtain the intermediate quality index Y 1_1 , through the intermediate function f 1_2 (Δt) Calculate the sampling data to obtain the intermediate quality index Y 1_2 , through the intermediate function f 1_3 (Δt) Calculate the sampling data to obtain the intermediate quality index Y 1_3 ; According to the intermediate quality index Y 1_1 , intermediate quality index Y 1_2 and intermediate quality index Y 1_3 The predicted quality index is calculated to obtain the difference index corresponding to each intermediate quality index; the intermediate function that wins the bet on the minimum difference index is used to update the process time function.

[0080] In some embodiments, if the process time function f0(Δt) and the process time function f2(Δt) belong to the same process scenario, and the scene difference index corresponding to the process scenario is greater than or equal to the scene index threshold, the same scene functions f0(Δt) and f2(Δt) are updated synchronously, that is, Table 2 is updated to obtain a new alternative quality index as shown in Table 4.

[0081] Table 4

[0082]

[0083] In some embodiments, after the process time function is updated, the process time function in Table 2 is updated, and the alternative quality indicator Y in Table 2 is updated to obtain an updated Table 2. The obtained Table 4 is the updated Table 2, and then Table 3 is updated based on the updated Table 4.

[0084] In some embodiments, the embodiments of the present disclosure provide a method for aligning quality indicators between process nodes, and the specific steps include: Step 1: Set the process duration function, calculate the alternative quality indicator Y, and give an expression from X to Y. Step 2: Train the expression G(X) from X to Y, and then calculate the predicted quality indicator Y'=G(X) at each time point. Step 3: Calculate the MSE for all the alternative quality indicators Y and the predicted quality indicators Y'; Step 4: If the MSE meets the requirements or the number of iterations reaches the upper limit, the predicted indicator Y' is the final quality indicator, otherwise, update the process duration function of each category according to the rules and return to step 1, wherein the smaller the difference indicator, the better.

[0085] Optionally, the method further includes: obtaining a current process flow, wherein the current process flow includes a plurality of original process nodes arranged in a preset process sequence, and an indicator alignment model is set between any two original process nodes; determining the current process flow as a test process; in response to the test process, determining the original process node with the first order in the test process as the test initial node, determining the original process node with the last order in the test process as the test termination node, determining the indicator alignment model between two adjacent original process nodes in the test process as the first alignment model, and determining the indicator alignment model between the test initial node and the test termination node as the second alignment model; obtaining a quality indicator sample corresponding to the test initial node; and using each first alignment model in turn to obtain the quality indicator sample at the test termination node. The first test indicator corresponding to the point is obtained, and the second alignment model is used to obtain the second test indicator corresponding to the quality indicator sample at the test termination node; the difference is calculated according to the first test indicator and the second test indicator to obtain the model normal indicator; if the model normal indicator is less than the preset normal indicator threshold, the test process is adjusted by the preset adjustment strategy to obtain a new test process, wherein the preset adjustment strategy includes deleting the test initial node from the test process, or deleting the test termination node from the test process; if the model normal indicator is greater than or equal to the normal indicator threshold, the abnormal model is determined from the indicator alignment model according to the comparison result between the current process flow and the test process, and the abnormal model is updated, wherein the abnormal model belongs to the current process flow, and the abnormal model does not belong to the test process.

[0086] In some embodiments, the current process flow consists of multiple original process nodes arranged in a preset process order, and these nodes are: raw material inspection (node ​​A), cutting (node ​​B), welding (node ​​C), polishing (node ​​D), quality inspection (node ​​E), wherein an indicator alignment model is set between any two original process nodes to ensure that the quality indicators between the nodes are consistent; the current process flow is determined as a test flow, i.e. A→B→C→D→E; the first-ranked original process node A in the test flow is determined as the test initial node, and the last-ranked original process node E in the test flow is determined as the test termination node; the indicator alignment model between two adjacent original process nodes in the test flow is determined as the first alignment model, i.e. A→B (first alignment model 1), B→C (first alignment model 2), C→D (first alignment model 3), and D→E (first alignment model 4). The index alignment model between the test initial node A and the test termination node E is determined as the second alignment model (global alignment model); the corresponding quality index sample is obtained from the test initial node A, and the sample includes various quality parameters of the raw materials; the quality index sample is gradually converted from the test initial node A to the test termination node E using each first alignment model (alignment model 1 to alignment model 4) in turn to obtain the first test index, and at the same time, the quality index sample is directly converted to the test termination node E using the second alignment model (global alignment model) to obtain the second test index; the difference calculation is performed according to the first test index and the second test index to obtain the model normal index. This index reflects the degree of consistency between each first alignment model and the global alignment model; if the model normal index is less than the preset normal index threshold (for example, the threshold is set to 5%), it indicates that there is a process node or the alignment model may be unreasonable or inefficient; in this example, it is assumed that the model normal index is 4%, which is less than the threshold, so the process adjustment is required; according to the preset adjustment strategy, the test process is adjusted. The preset adjustment strategy includes deleting the test initial node or the test termination node from the test process. In this example, considering that raw material inspection (node ​​A) is the basic link of product quality control, it is decided to delete the test termination node E (quality inspection) and integrate the quality inspection link into the previous process node (polishing D) to form a new test process A→B→C→D (after optimization); until the difference between the first test indicator and the second test indicator in the test process meets the model requirements, it means that there is an anomaly in the first alignment model outside the test process, and the process time function in the abnormal model is updated.

[0087] Combination Figure 6 As shown, the embodiment of the present disclosure provides a method for aligning quality indicators between process nodes, including:

[0088] Step S601, establish an expert knowledge base for the current process flow, and collect historical process data corresponding to the current process flow, and jump to step S602 and step S604;

[0089] Among them, the current process flow includes an initial process node and a target process node;

[0090] Step S602, determining a process duration function by combining an expert knowledge base and historical process data;

[0091] Step S603, calculating the sampled data according to the process duration function to obtain the candidate quality index corresponding to the target process node, and then jumping to step S606;

[0092] Step S604, determining an indicator prediction function in combination with an expert knowledge base, and performing model training on an original prediction model generated based on the indicator prediction function according to training samples with sample labels to obtain an indicator prediction model;

[0093] Among them, raw material quality indicators and historical process data associated with raw material quality indicators are used as training samples;

[0094] Among them, the product quality index associated with the raw material quality index is used as the sample label of the training sample;

[0095] Step S605, predicting using the indicator prediction model according to the initial quality indicator to obtain the predicted quality indicator corresponding to the target process node, and then jumping to step S606;

[0096] Step S606, determining a difference index according to the candidate quality index and the predicted quality index;

[0097] Step S607, determining whether the difference index meets the difference index threshold, if yes, jump to step S608, if no, jump to step S609;

[0098] Step S608: Outputting the candidate quality indicator through the indicator alignment model as the target quality indicator to which the initial quality indicator is aligned.

[0099] Step S609, update the process duration function and jump to step S603.

[0100] The method for aligning quality indicators between process nodes provided by the embodiment of the present disclosure is used to calculate the sampling data between the initial process node and the target process node through the process duration function to obtain the alternative quality indicator, and the indicator prediction model is used to predict the initial quality indicator corresponding to the initial process node to obtain the predicted quality indicator, so as to update the process duration function based on the difference between the alternative quality indicator and the predicted quality indicator, so as to output the target quality indicator corresponding to the initial quality indicator according to the updated alignment model, which has the following advantages:

[0101] First, the quality index of the target process node is assumed based on the alternative quality index obtained from the process duration function and the sampling data, and the predicted quality index obtained from the index prediction model is used as a reference for the alignment index, so that the difference between the alternative quality index and the predicted quality index is eliminated by updating the process duration function, and the updated index is aligned with the model to obtain the aligned target quality index, thereby greatly improving the alignment accuracy of the quality index and accurately reflecting the production status of the process node;

[0102] Second, the process time function is determined through the expert knowledge base to ensure that the assumed error of the quality indicator is within a reasonable range, and the indicator prediction model is obtained through the machine learning algorithm, so as to generate an iterative calculation formula based on indicator assumptions and indicator predictions according to the process time function and the indicator prediction model, so as to match the corresponding target quality indicator according to the initial quality indicator, thereby greatly improving the alignment accuracy of the quality indicator.

[0103] Combination Figure 7 As shown, the embodiment of the present disclosure provides a system for aligning quality indicators between process nodes, including a setting module 701 , an acquisition module 702 , an update module 703 and an output module 704 .

[0104] The setting module 701 is used to set an indicator alignment model between the initial process node and the target process node. The indicator alignment model includes a process duration function and an indicator prediction model, wherein the process duration function is used to characterize the corresponding relationship between the process duration and the probability density.

[0105] The acquisition module 702 is used to acquire the initial quality index corresponding to the initial process node, and perform quality detection in the time period between the initial process node and the target process node to obtain sampling data.

[0106] The update module 703 is used to make predictions based on the initial quality indicators using the indicator prediction model to obtain the predicted quality indicators corresponding to the target process node, and to calculate the sampled data based on the process duration function to obtain the alternative quality indicators corresponding to the target process node, so as to update the process duration function based on the difference between the predicted quality indicators and the alternative quality indicators.

[0107] The output module 704 is used to output the candidate quality indicator through the updated indicator alignment model as the target quality indicator aligned with the initial quality indicator.

[0108] The method for aligning quality indicators between process nodes provided by the embodiment of the present disclosure is adopted. The sampling data between the initial process node and the target process node is calculated through the process duration function to obtain the alternative quality indicator, and the indicator prediction model is used to predict the initial quality indicator corresponding to the initial process node to obtain the predicted quality indicator, so as to update the process duration function based on the difference between the alternative quality indicator and the predicted quality indicator, thereby outputting the target quality indicator corresponding to the initial quality indicator according to the updated alignment model. In this way, the quality indicator of the target process node is assumed according to the alternative quality indicator obtained from the process duration function and the sampling data, and the predicted quality indicator obtained according to the indicator prediction model is used as a reference for the alignment indicator, so as to eliminate the difference between the alternative quality indicator and the predicted quality indicator by updating the process duration function, and the updated indicator alignment model is used to obtain the aligned target quality indicator, thereby greatly improving the alignment accuracy of the quality indicator and accurately reflecting the production status of the process node.

[0109] An embodiment of the present disclosure further provides an electronic device, including: a processor and a memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device executes the above method.

[0110] Figure 8 The structure diagram of the computer system suitable for implementing the electronic device of the embodiment of the present application is shown. It should be noted that: Figure 8 The computer system 800 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0111] like Figure 8 As shown, the computer system 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 802 or the program loaded from the storage part 808 to the random access memory (RAM) 803, such as executing the method in the above embodiment. In the RAM 803, various programs and data required for system operation are also stored. The CPU 801, ROM 802 and RAM 803 are connected to each other through a bus 804. The input / output (I / O) interface 805 is also connected to the bus 804.

[0112] The following components are connected to the I / O interface 805: an input section 806 including a keyboard, a mouse, etc.; an output section 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed so that a computer program read therefrom is installed into the storage section 808 as needed.

[0113] The electronic device disclosed in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run the computer program so that the electronic device executes each step of the above method.

[0114] The above description and the accompanying drawings fully illustrate the embodiments of the present disclosure so that those skilled in the art can practice them. Other embodiments may include structural, logical, electrical, process and other changes. The embodiments represent possible changes only. Unless explicitly required, separate components and functions are optional, and the order of operation can be changed. Parts and sub-samples of some embodiments may be included in or replace parts and sub-samples of other embodiments. Moreover, the words used in this application are only used to describe the embodiments and are not used to limit the claims. As used in the description of the embodiments and the claims, unless the context clearly indicates, the singular forms of "a", "an" and "the" are intended to include plural forms as well. Similarly, the term "and / or" as used in this application refers to any and all possible combinations of listings containing one or more associated ones. In addition, when used in the present application, the term "comprise" and its variants "comprises" and / or comprising refer to the presence of a stated sub-sample, whole, step, operation, element, and / or component, but do not exclude the presence or addition of one or more other sub-samples, wholes, steps, operations, elements, components and / or groups of these. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the presence of other identical elements in the process, method or device comprising the elements. In this article, each embodiment may focus on the differences from other embodiments, and the same and similar parts between the various embodiments may refer to each other. For the methods, products, etc. disclosed in the embodiments, if they correspond to the method part disclosed in the embodiments, then the relevant parts can refer to the description of the method part.

[0115] Those skilled in the art will appreciate that the units and algorithm steps of each example 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 performed in hardware or software may depend on the specific application and design constraints of the technical solution. Technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the embodiments of the present disclosure. Technicians can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0116] In the embodiments disclosed herein, the disclosed methods and products (including but not limited to devices, equipment, etc.) can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units can be only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some sub-samples can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to implement this embodiment. In addition, each functional unit in the embodiment of the present disclosure may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0117] The flowchart and block diagram in the accompanying drawings show the possible architecture, functions and operations of the system, method and computer program product according to the embodiment of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and a part of the module, program segment or code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

Claims

1. A method for aligning quality indicators between process nodes, characterized in that: include: Presetting an indicator alignment model between an initial process node and a target process node, the indicator alignment model comprising a process duration function and an indicator prediction model, wherein the process duration function is used to characterize a corresponding relationship between process duration and probability density; Acquire an initial quality indicator corresponding to the initial process node, and perform quality inspection in a time period between the initial process node and the target process node to obtain sampling data; Predicting using the indicator prediction model according to the initial quality indicator to obtain a predicted quality indicator corresponding to the target process node, and calculating the sampled data according to the process duration function to obtain an alternative quality indicator corresponding to the target process node, so as to update the process duration function according to the difference between the predicted quality indicator and the alternative quality indicator; The updated indicator alignment model outputs an alternative quality indicator as a target quality indicator aligned with the initial quality indicator.

2. The method according to claim 1, characterized in that The process duration function is obtained by: Establishing an expert knowledge base for the current process flow in advance, and collecting historical process data corresponding to the current process flow, wherein the current process flow includes the initial process node and the target process node; Generate a process duration library in combination with the expert knowledge base and / or the historical process data, wherein the process duration library stores a plurality of process scenario data and original duration functions corresponding to each of the process scenario data; The current scene data is acquired, and matching is performed from the process duration library according to the matching result between the current scene data and each of the process scene data to obtain a process duration function.

3. The method according to claim 2, characterized in that The alternative quality index is determined by the following formula: Y = ∑y(t+Δt)×f(Δt); In the formula, Y is the alternative quality indicator, y(t+Δt) is the sampling data, f(Δt) is the process time function, t is the initial time point corresponding to the initial process node, and Δt is the process time between the initial process node and the target process node.

4. The method according to claim 2, characterized in that: Obtain the indicator prediction model in the following ways: Obtaining a raw material quality index corresponding to the initial process node, and obtaining a product quality index corresponding to the target process node, wherein there is a correlation relationship between the historical process data, the raw material quality index, and the product quality index; Establishing an index prediction function between the raw material quality index and the product quality index in combination with the expert knowledge base; The raw material quality index and the historical process data associated with the raw material quality index are used as training samples, and the product quality index associated with the raw material quality index is used as the sample label of the training sample. The original prediction model generated based on the indicator prediction function is trained according to the training samples with sample labels, and the trained original prediction model is used as the indicator prediction model.

5. The method according to claim 1, characterized in that The process duration function is updated according to the difference between the predicted quality indicator and the candidate quality indicator, including: If there are multiple process time functions, then the predicted quality indicators are calculated with the corresponding candidate quality indicators to obtain independent difference indicators corresponding to the candidate quality indicators, and the global difference indicator is calculated according to the independent difference indicators; If the global difference index is greater than or equal to a preset global index threshold, the process duration functions are classified according to the process scenario to obtain the same-scenario functions corresponding to different process scenarios, wherein the same-scenario functions corresponding to the same process scenario are the same; Taking any process scenario as a target scenario, calculating the predicted quality index and the target quality index corresponding to the same target scenario, and obtaining the scenario difference index corresponding to the target scenario; If the scene difference index is greater than or equal to a preset scene index threshold, each same-scene function corresponding to the target scene is updated; If all the process duration functions are updated, the update times of the process duration functions are counted, and the global difference index is updated according to the updated process duration functions; If the global difference index is less than the global index threshold, or the number of updates is greater than or equal to a preset number threshold, the model update state of the index alignment model is determined as update completion.

6. The method according to claim 5, characterized in that The independent variance index is calculated as follows: Obtain one or more difference index types, wherein the difference index types include one or more of mean square error, root mean square error and determination coefficient; The prediction quality indicator and the target quality indicator are calculated respectively according to each difference indicator type to obtain independent difference indicators corresponding to each difference indicator type.

7. The method according to claim 6, characterized in that The same-scene functions corresponding to the target scene are updated, including: Generate multiple intermediate functions corresponding to any one of the same scene functions by adjusting function parameters, wherein the interval length corresponding to the process duration interval in the intermediate function is smaller than the interval length corresponding to the process duration interval in the same scene function; Substituting each of the intermediate functions into each of the same-scene functions corresponding to the target scene, respectively, to calculate the scene difference index corresponding to the target scene, and obtaining the intermediate difference index corresponding to each of the intermediate functions; The objective function is determined from each of the intermediate functions according to the comparison results between the intermediate difference indicators, and the same-scene function is updated according to the objective function.

8. The method according to any one of claims 1 to 7, characterized in that: The method further comprises: Acquire a current process flow, wherein the current process flow includes a plurality of original process nodes arranged in a preset process sequence, and an indicator alignment model is set between any two original process nodes; Determine the current process flow as a test process; In response to a test flow, determining the original process node with the first order in the test flow as a test initial node, determining the original process node with the last order in the test flow as a test termination node, determining the indicator alignment model between two adjacent original process nodes in the test flow as a first alignment model, and determining the indicator alignment model between the test initial node and the test termination node as a second alignment model; Obtaining a quality indicator sample corresponding to the initial node of the test; Using the first alignment models in sequence, obtain the first test indicator corresponding to the quality indicator sample at the test termination node, and using the second alignment model, obtain the second test indicator corresponding to the quality indicator sample at the test termination node; Perform difference calculation based on the first test index and the second test index to obtain a model normality index; If the normal index of the model is less than a preset normal index threshold, the test process is adjusted by a preset adjustment strategy to obtain a new test process, wherein the preset adjustment strategy includes deleting the test initial node from the test process, or deleting the test termination node from the test process; If the normal indicator of the model is greater than or equal to the normal indicator threshold, the abnormal model is determined from the indicator alignment model according to the comparison result between the current process flow and the test flow, and the abnormal model is updated, wherein the abnormal model belongs to the current process flow and the abnormal model does not belong to the test flow.

9. A quality indicator alignment system for process nodes, characterized in that: include: A setting module, used to set an indicator alignment model between an initial process node and a target process node, wherein the indicator alignment model includes a process duration function and an indicator prediction model, wherein the process duration function is used to characterize a corresponding relationship between process duration and probability density; An acquisition module, used to acquire an initial quality indicator corresponding to the initial process node, and perform quality detection in a time period between the initial process node and the target process node to obtain sampling data; An updating module, configured to predict the initial quality indicator using the indicator prediction model to obtain the predicted quality indicator corresponding to the target process node, and calculate the sampled data according to the process duration function to obtain the alternative quality indicator corresponding to the target process node, so as to update the process duration function according to the difference between the predicted quality indicator and the alternative quality indicator; The output module is used to output the alternative quality indicator through the updated indicator alignment model as the target quality indicator aligned with the initial quality indicator.

10. An electronic device, characterized in that: include: Processor and memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device executes the method according to any one of claims 1 to 8.

Images