CN113361661A - Modeling method and device for data cooperation capability evaluation - Google Patents

Abstract

The present application discloses a modeling method and device for evaluating data collaboration ability. The modeling method includes: constructing a variable acquisition model, and the variable acquisition model is used to obtain the values of multiple variables according to the collected process data. The variables include process precision and Process accuracy; build a variable ability evaluation model, which is used to calculate the ability evaluation score of the corresponding variable based on the value of the variable; build a collaborative ability acquisition model, which is used to calculate the ability evaluation score based on multiple variables Data collaboration ability score; the variable acquisition model, the variable ability evaluation model and the collaboration ability acquisition model are combined to form a data collaboration ability evaluation pre-training model; the data collaboration ability evaluation pre-training model is trained to obtain a data collaboration ability evaluation model. This application uses multiple variables to evaluate the data synergy capability, and objectively, scientifically and accurately describes the data synergy capability in the production process.

Description

A modeling method and device for data collaboration capability evaluation

technical field

The present application relates to the technical field of production and manufacturing, and more particularly, to a modeling method and device for evaluating data collaboration capability.

Background technique

The data synergy capability of cigarette shredding products refers to the degree to which the products of the previous process can meet the processing requirements of the next process, such as: inlet moisture content, outlet temperature, outlet moisture content, etc. In the process of cross-process production, many factors such as the properties of incoming materials, ambient temperature and humidity, and the way of cloth in the cabinet will affect the product. In view of cross-workshop data collaboration, in order to meet the quality requirements of wrapping and rolling, the whole silk rate, broken silk rate, moisture content, filling value, temperature and other indicators of each brand of finished tobacco must be stably controlled, and the stability of these indicators must be controlled. Relying on the stability, control and synergy of the processing parameters during the silk making process, the homogenization of the product quality can be achieved.

The research on the data synergy capability of silk-making products is a complex subject, and improving the data synergy capability is a key part of improving the quality of finished cut tobacco. In the production process of shredding, tobacco leaves are processed from raw materials into finished tobacco, which has completed multiple cross-process production. In each cross-procedure and cross-departmental production process, collaborative production data is generated. At present, in the tobacco industry, there are few evaluations of the data synergy ability of silk-making products, most of which are limited to the traditional single-process quality index management, and lack of comprehensive evaluation of more production links.

At present, the evaluation of the data synergy ability of silk-making products is mainly aimed at the production link of "loose moisture regain- _feeding -leaf storage-drying". A simple assessment of moisture can complete the evaluation of product data synergy ability. This evaluation method only considers the process control ability of a single process or single characteristic, the evaluation scope is small and not precise enough, and the value of collaborative data cannot be tapped. There is no coordination problem, and there is a lack of evaluation of cross-departmental data coordination capabilities, which is not conducive to improving the coordination capabilities of production data and the level of product homogeneity.

Moreover, when both the process precision C _p and the process accuracy C _a have large co-directional deviations, the calculation process of the existing process capability index C _pk will make these two deviations cancel each other to a certain extent, so that the process cannot be correctly reflected. unstable state.

SUMMARY OF THE INVENTION

The present application provides a modeling method and device for evaluating the data collaboration capability, which uses multiple variables to evaluate the data collaboration capability and objectively, scientifically and accurately describes the data collaboration capability in the production process.

This application provides a modeling method for data collaboration capability evaluation, including:

Build a variable acquisition model, the variable acquisition model is used to obtain the values of multiple variables according to the collected process data, and the variables include process precision and process accuracy;

Build an ability evaluation model of the variable, and the ability evaluation model is used to calculate the ability evaluation score of the corresponding variable according to the value of the variable;

Build a collaborative ability acquisition model, and the collaborative ability acquisition model is used to calculate the data collaborative ability score based on the ability evaluation scores of multiple variables;

The variable acquisition model, the variable ability evaluation model and the collaborative ability acquisition model are combined to form a data collaborative ability evaluation pre-training model;

The data collaboration capability evaluation pre-training model is trained to obtain the data collaboration capability evaluation model.

Preferably, constructing the variable acquisition model includes preprocessing the collected process data to obtain the preprocessed process data.

Preferably, constructing the variable acquisition model further includes taking the preprocessed process data as a sample, calculating the standard deviation and mean of the sample, and further calculating the values of multiple variables of the sample.

Preferably, the ability evaluation model for constructing variables includes:

Obtain the membership function and its parameters corresponding to each variable through fuzzy algorithm;

According to the parameters of the membership function and the value of the variable, the ability evaluation score of the corresponding variable is calculated.

Preferably, the capability evaluation score of the variable is calculated using a percentile capability index function.

Preferably, the following formula is used to calculate the data synergy ability score F

Among them, f ₁ represents the capability evaluation score of process precision, f ₂ represents the capability evaluation score of process accuracy, and α represents the weight of process precision.

The present application also provides a modeling device for data collaboration capability evaluation, including a variable acquisition model construction module, a variable capability evaluation model construction module, a collaborative capability acquisition model construction module, a pre-training model construction module, and a training module;

The variable acquisition model building module acquires the values of multiple variables according to the collected process data, and the variables include process precision and process accuracy;

The ability evaluation model building module of the variable is used to calculate the ability evaluation score of the corresponding variable according to the value of the variable;

The collaborative ability acquisition model building module is used to calculate the data collaborative ability score according to the ability evaluation scores of multiple variables;

The pre-training model building module is used to combine the variable acquisition model, the variable ability evaluation model and the collaborative ability acquisition model to form a data collaborative ability evaluation pre-training model;

The training module trains the data collaboration ability evaluation pre-training model to obtain the data collaboration ability evaluation model.

Preferably, the variable acquisition model building module includes a data preprocessing submodule, and the data preprocessing submodule is used to preprocess the collected process data to obtain the preprocessed process data.

Preferably, the variable ability evaluation model building module includes a fuzzy algorithm sub-module and an ability evaluation score calculation sub-module;

The fuzzy algorithm sub-module is used to obtain the membership function and its parameters corresponding to each variable through the fuzzy algorithm;

The ability evaluation score calculation sub-module is used to calculate the ability evaluation score of the corresponding variable according to the parameter of the membership function and the value of the variable.

Preferably, the ability evaluation score calculation module calculates the ability evaluation score of the variable by using a percentile system ability index function.

Other features and advantages of the present application will become apparent from the following detailed description of exemplary embodiments of the present application with reference to the accompanying drawings.

Description of drawings

The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application.

Fig. 1 is the flow chart of the modeling method of data collaboration ability evaluation provided by this application;

FIG. 2 is a structural diagram of a modeling apparatus for data collaboration capability evaluation provided by the present application.

Detailed ways

Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of the components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present application unless specifically stated otherwise.

The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the application, its application, or uses.

Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification.

In all examples shown and discussed herein, any specific values should be construed as illustrative only and not limiting. Accordingly, other instances of the exemplary embodiment may have different values.

It should be noted that this application is applicable to the data synergy capability of cigarette shredding products, as well as the cross-process data synergy capability evaluation of stem products, the cross-department data synergy capability evaluation from the production department to the operation department, and the production of other products. process, realize the transparency of production data, and provide data support for product quality management.

FIG. 1 is a flow chart of the modeling method for data collaboration capability evaluation provided by the present application. As shown in Figure 1, the modeling method for data collaboration capability evaluation includes the following steps:

S110: Build a variable acquisition model, which is used to acquire values of multiple variables according to the collected process data, where the variables include process precision C _p , process accuracy C _a and process capability index C _pk . Among them, the process precision C _p and the process accuracy C _a are used to calculate the data synergy score, and the process capability index C _pk is used to verify the model.

Process precision is a measure of the degree to which process control meets product quality standards. The larger the C _p , the smaller the variation and the better the process control capability. Process control accuracy is an index to measure the consistency between the actual center value and the standard center value in process control. The larger the C _a , the greater the deviation between the actual center value and the specification center value, and the worse the process control ability.

Building a variable acquisition model includes taking the process data as a sample, calculating the standard deviation σ and the mean μ of the sample, and further calculating the values of multiple variables of the sample.

in,

T=USL-LSL (3)

C _pk =(1-|Ca|)·C _p (4)

where USL is the upper standard limit, LSL is the lower standard limit, M is the standard median, and |Ca| is the absolute value of the value of the process accuracy.

Preferably, the collected process data is preprocessed before the variables are calculated to obtain preprocessed process data. Data preprocessing includes at least data cleaning and data interception.

Among them, data cleaning includes but is not limited to: effective data screening of the raw process data extracted from the Manufacturing Execution System (MES); filling in missing values using the hot card imputation method; identifying and eliminating outliers using the 3σ principle : Noise detection is performed on the data, and the data is smoothed by the binning method to remove noise. As an embodiment, in a cross-process or cross-department production link, the process data of the last process or department of the production link is collected as the process raw data in step S110. For example, in the production link of "loose and moisturizing-feeding-leaf storage-silk drying", the raw process data of the "silk drying" process are collected.

Data interception includes, but is not limited to, identifying steady-state and non-steady-state process data, and intercepting steady-state data as a subsequent data basis.

S120: Build an ability evaluation model of the variable, and the ability evaluation model is used to calculate the ability evaluation score of the corresponding variable according to the value of the variable.

Specifically, as an embodiment, constructing a variable ability evaluation model includes:

S1201: Obtain a membership function and its parameters corresponding to each variable through a fuzzy algorithm.

Specifically, a larger or smaller distribution is selected as the membership function of the fuzzy set in the fuzzy algorithm, and the strategy of the optimal membership function is used to select the final membership function, and the optimal value of each parameter in the membership function is fitted. The value of the variables of the evaluation model is set to a unified and standardized judgment standard.

S1202: Calculate the ability evaluation score of the corresponding variable according to the parameter of the membership function and the value of the variable.

Specifically, as an embodiment, the capability evaluation score of the variable is calculated by using the capability index function of the percentile system, and the capability index of the variable is standardized and calculated by the percentile system.

The constructed percentile capability index function is as follows:

Among them, f ₁ represents the ability evaluation score of the process precision C _p , x represents the value of C _p , a, b are the parameters of the membership function corresponding to the process precision;

_{f 2} represents the capability evaluation score of the process accuracy Ca, z represents the value of |Ca|, and c and d are the parameters of the membership function corresponding to the process accuracy.

S130: Build a synergy ability acquisition model, and the synergy ability acquisition model is used to calculate the data synergy ability score according to the ability evaluation scores of multiple variables.

Specifically, the following formula is used to calculate the data collaboration ability score F

Among them, f ₁ represents the capability evaluation score of process precision, f ₂ represents the capability evaluation score of process accuracy, and α represents the weight of process precision.

It should be noted that the determination of the weight α is mainly based on the control requirements of each process or department, combined with the degree of influence of the two variables in the process or department on the sensory quality, and the Delphi method is used to evaluate the process precision. Cross-process (for example, between multiple processes in the production link of "loose and moisturizing-feeding-leaf storage-drying", and between multiple processes in the production link of "drying-winding-fragrance"), cross-departmental (For example, among multiple departments in the production link of "stereo wire library - wind-driven wire feeding package") the production weight α is revised. Taking the cross-process of "loose and moisturizing-feeding-leaf storage-drying" as an example, the value of its weight α is set as 0.4 after research.

S140: Combine the variable acquisition model, the variable ability evaluation model, and the collaboration ability acquisition model to form a data collaboration ability evaluation pre-training model.

S150: Train the pre-training model for data collaboration capability evaluation to obtain a data collaboration capability evaluation model.

When verifying the data collaboration capability evaluation model, as an embodiment, two methods are adopted: dynamic characterization performance verification and sensitive characterization performance verification. In the dynamic characterization performance verification, the correlation analysis is carried out on the process capability index C _pk and the data synergy capability score. In the verification of sensitive characterization performance, it is verified whether the data synergy score can characterize the comprehensive fluctuation of process precision C _p and process accuracy _Ca .

The verification standard for the evaluation model of data synergy ability is: the correlation coefficient in the dynamic characterization performance verification is greater than or equal to 0.8, and the sensitive characterization performance verification results are passed. If the above criteria are met, it is judged that the evaluation of the data synergy evaluation model is accurate and the evaluation effect is excellent.

Taking the evaluation of cross-process data synergy ability of "loose and moisturizing-feeding-leaf storage-drying" as an example, the effect of the evaluation model of data synergy ability was verified. In the dynamic characterization performance verification, the correlation coefficient between the data synergy score and the process capability index C _pk is 0.975, indicating that the results of the data synergy evaluation model can objectively characterize the level of production parameter synergy. In the verification of sensitive characterization performance, a part of batch data with special characteristics is called for verification calculation. It can be seen from the calculation results that for two different batches with the same process capability index C _pk , the data synergy ability scores have different degrees. It shows that the data synergy score can more realistically characterize the deviation and discrete comprehensive fluctuation of production parameters in process control, and can still characterize different processes when the process precision C _p and the process accuracy C _a cancel each other out. difference. Therefore, the data synergy evaluation method of the present application can more objectively, scientifically and accurately characterize the performance of the data synergy capability of silk-making products, and the model is judged to be accurate in evaluation and excellent in evaluation effect.

When using the trained model to evaluate the data synergy ability, input the collected process data into the data synergy ability evaluation model to obtain the data synergy ability score, and determine whether it meets the standard according to the level to which the data synergy ability score belongs.

As an example, the standard table for the level of collaboration capability is as follows:

Table 2 Classification standard table of synergy ability level

Based on the above-mentioned modeling method for data collaboration capability evaluation, the present application further provides a modeling device for data collaboration capability evaluation. As shown in FIG. 2 , the modeling apparatus includes a variable acquisition model construction module 210 , a variable capability evaluation model construction module 220 , a collaborative capability acquisition model construction module 230 , a pretrained model construction module 240 and a training module 250 .

The variable acquisition model building module 210 acquires the values of a plurality of variables according to the collected process data, and the variables include process precision and process accuracy.

Preferably, the variable acquisition model building module includes a data preprocessing module and a variable calculation submodule. The data preprocessing sub-module is used to preprocess the collected process data to obtain the preprocessed process data. The variable calculation submodule is used to take the preprocessed process data as a sample, calculate the standard deviation and mean of the sample, and further calculate the values of multiple variables of the sample.

The ability evaluation model building module 220 of the variable is used to calculate the ability evaluation score of the corresponding variable according to the value of the variable.

As an embodiment, the variable ability evaluation model building module includes a fuzzy algorithm sub-module and an ability evaluation score calculation sub-module.

The fuzzy algorithm sub-module is used to obtain the membership function and its parameters corresponding to each variable through the fuzzy algorithm.

The ability evaluation score calculation sub-module is used to calculate the ability evaluation score of the corresponding variable according to the parameter of the membership function and the value of the variable.

The synergy capability acquisition model building module 230 is configured to calculate the data synergy capability score according to the capability evaluation scores of multiple variables.

The pre-training model building module 240 is configured to combine the variable acquisition model, the variable ability evaluation model and the collaborative ability acquisition model to form a data collaborative ability evaluation pre-training model.

The training module 250 trains the data coordination ability evaluation pre-training model to obtain the data coordination ability evaluation model.

The beneficial effects obtained by this application are as follows:

1. This application uses multiple variables to evaluate the data synergy capability, and objectively, scientifically and accurately describes the data synergy capability in the production process.

2. Compared with the method in the prior art that only uses a single index C _pk to evaluate the data synergy capability, the evaluation method in the present application is conducive to eliminating the situation that the process accuracy and process precision occur in the same direction deviation, and the evaluation result obtained. More scientific and accurate.

3. The capability evaluation score of the variables in this application is a percentage system, and the numerical resolution is higher than the process capability index C _pk , which can be directly used as a performance indicator.

4. The data collaboration capability evaluation model of this application automatically generates a data collaboration capability analysis report, which effectively improves the quality of cigarette shredded products. The operator can directly understand the cross-process and cross-departmental performance of the product within a certain period of time according to the analysis report, and quickly Find batches with insufficient synergy ability, and effectively improve the quality of cigarette shreds while improving work efficiency.

Although some specific embodiments of the present application have been described in detail by way of examples, those skilled in the art should understand that the above examples are for illustrative purposes only and are not intended to limit the scope of the present application. Those skilled in the art will appreciate that modifications may be made to the above embodiments without departing from the scope and spirit of the present application. The scope of the application is defined by the appended claims.

Claims (10)

1. A modeling method for data synergy capability evaluation is characterized by comprising the following steps:

constructing a variable acquisition model, wherein the variable acquisition model is used for acquiring values of a plurality of variables according to acquired process data, and the variables comprise process precision and process accuracy;

constructing a capability evaluation model of the variable, wherein the capability evaluation model is used for calculating the capability evaluation score of the corresponding variable according to the value of the variable;

constructing a cooperation capability obtaining model, wherein the cooperation capability obtaining model is used for calculating a data cooperation capability score according to the capability evaluation scores of the variables;

combining the variable acquisition model, the capability evaluation model of the variable and the cooperation capability acquisition model to form a data cooperation capability evaluation pre-training model;

and training the data cooperation ability evaluation pre-training model to obtain a data cooperation ability evaluation model.

2. The modeling method for data synergy potential evaluation according to claim 1, wherein the constructing of the variable acquisition model comprises preprocessing the collected process data to obtain preprocessed process data.

3. The modeling method for data synergy potential evaluation according to claim 2, wherein the constructing a variable acquisition model further comprises taking the preprocessed process data as a sample, calculating a standard deviation and a mean of the sample, and further calculating values of a plurality of variables of the sample.

4. The modeling method for data synergy capability evaluation according to claim 1, wherein the constructing a capability evaluation model of variables comprises:

obtaining a membership function corresponding to each variable and parameters thereof through a fuzzy algorithm;

and calculating the capability evaluation score of the corresponding variable according to the parameters of the membership function and the value of the variable.

5. The method of claim 4, wherein the capability evaluation score of the variable is calculated using a percentile capability index function.

6. The modeling method for data interoperability evaluation according to claim 1, wherein the data interoperability score F is calculated using the following formula

Wherein f is₁Ability evaluation score, f, indicating process precision₂The ability evaluation score, which represents the accuracy of the process, and α represents the weight of the precision of the process.

7. A modeling device for data cooperation capability evaluation is characterized by comprising a variable acquisition model building module, a variable capability evaluation model building module, a cooperation capability acquisition model building module, a pre-training model building module and a training module;

the variable acquisition model construction module acquires values of a plurality of variables according to the acquired process data, wherein the variables comprise process precision and process accuracy;

the variable ability evaluation model building module is used for calculating the ability evaluation score of the corresponding variable according to the value of the variable;

the cooperation capability acquisition model construction module is used for calculating a data cooperation capability score according to the capability evaluation scores of the variables;

the pre-training model construction module is used for combining the variable acquisition model, the capability evaluation model of the variable and the cooperation capability acquisition model to form a data cooperation capability evaluation pre-training model;

and the training module trains the data cooperation ability evaluation pre-training model to obtain a data cooperation ability evaluation model.

8. The modeling apparatus for data synergy capability evaluation according to claim 7, wherein the variable acquisition model construction module includes a data preprocessing submodule, and the data preprocessing submodule is configured to preprocess the acquired process data to obtain preprocessed process data.

9. The modeling device for data synergy capability evaluation according to claim 7, wherein the capability evaluation model construction module for the variables comprises a fuzzy algorithm sub-module and a capability evaluation score calculation sub-module;

the fuzzy algorithm submodule is used for obtaining a membership function corresponding to each variable and parameters thereof through a fuzzy algorithm;

and the ability evaluation score calculation submodule is used for calculating the ability evaluation score of the corresponding variable according to the parameters of the membership function and the values of the variables.

10. The modeling apparatus for data collaboration capacity evaluation as claimed in claim 9, wherein the capacity evaluation score calculating module calculates the capacity evaluation score of the variable using a percentile capacity index function.

Images