Key index extraction method and system for digital manufacturing maturity evaluation

Abstract

The invention discloses a key index extraction method and system for digital manufacturing maturity evaluation, and relates to the technical field of data processing, systematic preprocessing and candidate index extraction are carried out on business data of each process of a workshop, invalid indexes are removed, and adaptive calibration is carried out, so that the evaluation accuracy of the manufacturing maturity is improved. Generating a first core candidate index set which can really reflect the digitization and automation level; and after grading and tolerance compensation are carried out based on the first core candidate index set, a finer second production unit evaluation grade is screened out again, and finally, a targeted optimization scheme is output for each process based on digital twin simulation and a rule base. According to the method, the accuracy of index extraction is greatly improved, the whole-process self-driven closed loop from candidate indexes to core indexes and then to evaluation grading and process optimization is realized through double-round grading and closed loop optimization, and a high-credibility quantitative basis and an operable improved path are provided for digital manufacturing maturity evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a key indicator extraction method and system for digital manufacturing maturity evaluation. BACKGROUND

[0002] Digital manufacturing maturity evaluation refers to the quantitative or qualitative classification evaluation of the development stage and capability level of manufacturing enterprises in the aspects of digitalization, automation and intelligentization through systematic methods and tools. The core goal is to help enterprises identify the current stage, advantages and disadvantages in the digital transformation path, thereby guiding subsequent resource investment and technology upgrading. As the basic link of the evaluation work, the main task of key indicator extraction is to identify the business characteristic quantities that best reflect the digitalization level and evolution trend from a large number of heterogeneous and dynamically changing business data in the manufacturing process.

[0003] For example, the publication number is: CN110298058A, a power grid supervision index key element extraction method and device is disclosed, which is based on the background of strengthening supervision of power grid enterprises under the new situation of power reform. By analyzing the power grid supervision files issued in recent years, the specific requirements of power grid enterprises are classified and analyzed. The influence of supervision requirements on power grid development is analyzed, and the influencing factors of power grid development are summarized. Based on ISM technology, the path of specific supervision requirements to power grid development is identified.

[0004] For example, the publication number is: CN114036784A, an aviation product research and development process and collaborative process modeling method based on maturity level is disclosed. The method includes: defining the aviation product maturity level based on the characterization elements and influencing elements of the full life cycle product maturity; establishing a parallel collaborative model for automatic control and design manufacturing of the aviation product development process based on the Petri net aviation product full life cycle model control scheme; analyzing and evaluating the performance and indicators of the aviation product life cycle process model to provide a basis for product parallel collaboration and shorten the research and development cycle.

[0005] However, in the process of implementing the technical scheme of the present application, the applicant found that the above-mentioned technology at least has the following technical problems: Firstly, the indicator extraction process lacks a dynamic response mechanism. Currently, the system often uses a fixed template type index system, which is difficult to adaptively update the indicators as the enterprise's specific business process, IT system architecture or workshop operation strategy changes, resulting in evaluation results lagging behind the actual digital development status. Secondly, the internal stability, discrimination, predictability and other evaluation characteristics of the indicators are not systematically analyzed, and they are rarely verified in depth from the dimensions of time series, data variability and statistical distribution, thereby affecting the interpretability of the indicators in the maturity level division. SUMMARY

[0006] In order to solve the technical problems existing in the prior art, the embodiment of the present application provides a key index extraction method for digital manufacturing maturity evaluation. The technical scheme is as follows: S1, obtaining and preprocessing intelligent factory workshop process business data, extracting process business data candidate index dataset, performing preliminary dynamic screening analysis on the process business data candidate index dataset, eliminating the candidate index dataset of the business data, and recalibrating the candidate index dataset of the business data to obtain a first core candidate index set.

[0007] S2, analyzing the first core candidate index set to obtain production unit evaluation factors of each core candidate index data in the first core candidate index set, screening to obtain a first production unit evaluation level of the business data, and performing tolerance compensation on the business data.

[0008] S3, re-screening the tolerance compensated business data to obtain a second production unit evaluation level grading result of the business data.

[0009] S4, adjusting and analyzing each process according to the second production unit evaluation level grading result of the business data to generate a candidate optimization scheme for each process.

[0010] Further, the process business data candidate index dataset is subjected to preliminary dynamic screening analysis, and the specific process is as follows: a sliding window is preset, the process business data candidate index dataset is extracted in the sliding window, and the completeness rate, extreme point quantity, median drift amount and drift mean value of the candidate index data of the process business data are obtained.

[0011] The completeness rate of the candidate index data of the process business data is extracted, and the defined completeness rate stored in the database, the extreme point quantity and the defined extreme point quantity stored in the database, the median drift amount and the defined median drift amount stored in the database, and the drift mean value and the defined drift mean value stored in the database are extracted, respectively. The weight coefficient is introduced to obtain the discrimination value of the business data candidate index dataset through proportion analysis, and the discrimination value of the business data candidate index dataset is used to quantitatively evaluate the effectiveness of the candidate index data of the process business data in distinguishing different manufacturing maturity stages.

[0012] Further, the candidate index dataset of the business data is eliminated, and the specific process is as follows: the discrimination value of the business data candidate index dataset is extracted, and compared with the set discrimination threshold value of the business data candidate index dataset. If the discrimination value of the business data candidate index dataset is lower than the discrimination threshold value of the business data candidate index dataset, the candidate index dataset of the business data is eliminated.

[0013] The discrimination value of the business data candidate indicator data set corresponding to each business data candidate indicator after the business data candidate indicator is removed is obtained in the candidate indicator data detection period, and the discrimination value of the business data candidate indicator data set corresponding to each business data candidate indicator after the business data candidate indicator is removed is subtracted from the discrimination threshold value of the business data candidate indicator data set to obtain the discrimination deviation value of the business data candidate indicator data set corresponding to each business data candidate indicator after the business data candidate indicator is removed. If the discrimination deviation value of the business data candidate indicator data set corresponding to each business data candidate indicator after the business data candidate indicator is removed is higher than the preset discrimination deviation threshold value of the business data candidate indicator data set, the removal of the business data candidate indicator is abandoned, and the indicator is re-added to the core candidate set. If the discrimination deviation value of the business data candidate indicator data set corresponding to each business data candidate indicator after the business data candidate indicator is removed is lower than or equal to the preset discrimination deviation threshold value of the business data candidate indicator data set, the business data candidate indicator is removed.

[0014] Further, the business data candidate indicator data set is recalibrated, and the specific process is: the discrimination value of the business data candidate indicator data set is extracted and compared with the set discrimination threshold value of the business data candidate indicator data set. If the discrimination value of the business data candidate indicator data set is lower than the discrimination threshold value of the business data candidate indicator data set, the business data candidate indicator data set is recalibrated to obtain the first core candidate indicator set. If the discrimination value of the business data candidate indicator data set is higher than or equal to the discrimination threshold value of the business data candidate indicator data set, the business data candidate indicator data set does not need to be recalibrated, and the first core candidate indicator set is directly obtained.

[0015] Further, the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set is obtained, and the specific process is: the value of each core candidate indicator data in the first core candidate indicator set is obtained, and the threshold value corresponding to each core candidate indicator data value stored in the database is extracted.

[0016] The average Pearson correlation coefficient of each core candidate indicator data and the target business volume stored in the database is obtained, which is recorded as the average Pearson correlation coefficient of each core candidate indicator data, and the Pearson correlation coefficient threshold value is extracted.

[0017] The data values of each core candidate indicator in the first core candidate indicator set are compared with the threshold value corresponding to the data value of each core candidate indicator, the average Pearson correlation coefficient of the data of each core candidate indicator is compared with the Pearson correlation coefficient threshold value corresponding to the average Pearson correlation coefficient, and the discrimination value and the weight coefficient of the business data candidate indicator data set are introduced to obtain the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set. The production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set is used to quantitatively evaluate the performance intensity of the production unit in each key dimension of digital manufacturing maturity.

[0018] Further, the business data first production unit evaluation level is obtained by screening, and the specific process is as follows: the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set is extracted and compared with the production unit evaluation factor interval stored in the database, the number of production unit evaluation factors of each core candidate indicator data in the first core candidate indicator set exceeding the production unit evaluation factor interval is counted, and is recorded as the number of out-of-limit indicators.

[0019] The number of out-of-limit indicators is extracted and compared with the unit evaluation level corresponding to each interval of the set number of out-of-limit indicators to obtain the business data first production unit evaluation level.

[0020] The business data first production unit evaluation level includes first, second and third levels.

[0021] Further, the business data is compensated for tolerance, and the specific process is as follows: the absolute value of the difference between the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set and the nearest boundary of the production unit evaluation factor interval is extracted, and the business data is compensated for tolerance according to the absolute value of the difference between the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set and the nearest boundary of the production unit evaluation factor interval.

[0022] Further, the business data after tolerance compensation is rescreened to obtain the business data second production unit evaluation level grading result, and the specific process is as follows: the number of out-of-limit indicators corresponding to the business data after tolerance compensation is counted, and if the number of out-of-limit indicators corresponding to the business data after tolerance compensation is higher than or equal to the set number of out-of-limit indicators corresponding to the business data after tolerance compensation threshold value, the business data first production unit evaluation level is adjusted downward, thereby obtaining the business data second production unit evaluation level grading result.

[0023] If the number of over-limit indicators corresponding to the tolerance-compensated business data is lower than the set threshold of the number of over-limit indicators corresponding to the tolerance-compensated business data, and the first production unit evaluation level of the business data is level one, the first production unit evaluation level of the business data is maintained; if the number of over-limit indicators corresponding to the tolerance-compensated business data is lower than the set threshold of the number of over-limit indicators corresponding to the tolerance-compensated business data, and the first production unit evaluation level of the business data is not level one, the first production unit evaluation level of the business data is adjusted upward, thereby obtaining the second production unit evaluation level classification result of the business data.

[0024] Further, according to the second production unit evaluation level classification result of the business data, each process is analyzed and adjusted to generate a candidate optimization scheme for each process. The specific process is as follows: extracting the second production unit evaluation level classification result of the business data, if the second production unit evaluation level classification result of the business data is not level one, calling a simulation model trained in advance in a digital twin environment, inputting process parameters into a virtual production line, and generating a candidate optimization scheme for each process.

[0025] On the other hand, the application also provides a key indicator extraction system for digital manufacturing maturity evaluation, comprising: a candidate indicator collection and calibration module for obtaining and preprocessing business data of each process in an intelligent factory workshop, extracting a candidate indicator data set of each process business data, performing preliminary dynamic screening analysis on the candidate indicator data set of each process business data, eliminating the candidate indicator data set of the business data, and recalibrating the candidate indicator data set of the business data to obtain a first core candidate indicator set.

[0026] A tolerance compensation module is configured to analyze the first core candidate indicator set, obtain a production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set, screen to obtain a first production unit evaluation level of the business data, and perform tolerance compensation on the business data.

[0027] A production unit level confirmation module is configured to rescreen the tolerance-compensated business data to obtain a second production unit evaluation level classification result of the business data.

[0028] An optimization process adjustment module is configured to analyze and adjust each process according to the second production unit evaluation level classification result of the business data, and generate a candidate optimization scheme for each process.

[0029] The technical scheme provided by the embodiment of the application has at least the following beneficial effects: (1) The application proposes a key index extraction method for digital manufacturing maturity evaluation. Through systematic preprocessing and candidate index extraction of the business data of each process in the workshop, invalid indexes are removed and self-adaptive calibration is performed to generate a first core candidate index set that can truly reflect the level of digitalization and automation. Based on the core index, hierarchical and tolerance compensation are performed, and a more refined second production unit evaluation grade is screened out. Finally, relying on digital twin simulation and rule base, a targeted optimization scheme is output for each process. Not only does it greatly improve the accuracy of index extraction, but also through double hierarchical classification and closed-loop optimization, it realizes the whole-process self-driven closed loop from candidate index to core index, to evaluation grading and process optimization, providing a high-credibility quantitative basis and an operable improvement path for digital manufacturing maturity evaluation.

[0030] (2) The application helps generate the first core candidate index set by obtaining the discrimination value of the business data candidate index data set, which helps ensure that the remaining indexes have sufficient stability and noise immunity, and also maximizes the retention of data dimensions that are most discriminative for manufacturing maturity evaluation, providing high-quality core input for subsequent hierarchical evaluation and process optimization.

[0031] (3) The application helps analyze the first production unit evaluation grade of the business data by obtaining the production unit evaluation factor of each core candidate index data in the first core candidate index set. Finally, according to the difference between each evaluation factor and its nearest interval boundary, tolerance compensation adjustment is automatically performed, which not only ensures the strictness of maturity grading, but also provides a buffer for slight deviations, so that the key index extraction not only accurately reflects the digitalization level, but also has the resilience and practical value against short-term fluctuations.

[0032] (4) The application helps call the simulation model that has been trained in the digital twin environment by using the second production unit evaluation grade classification result of the business data, imports the current process parameters into the virtual production line for rapid simulation, and generates a feasible process optimization scheme for each out-of-limit index. Through secondary classification, the robustness and accuracy of maturity evaluation are ensured, and with the help of digital twin closed loop, an implementable and verifiable improvement path is provided for continuous optimization of key indexes, thus realizing the whole-process self-driving from evaluation to optimization in the key index extraction link, significantly improving the effectiveness and guiding value of digital manufacturing maturity evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0034] Figure 1 is a flow chart of a key indicator extraction method for digital manufacturing maturity evaluation provided by an embodiment of the present application. Figure 2 is a logic flow diagram of a key indicator extraction method for digital manufacturing maturity evaluation provided by an embodiment of the present application. Figure 3 is a system module schematic diagram provided by an embodiment of the present application. DETAILED DESCRIPTION

[0035] The technical solutions in the present application will be described below with reference to the drawings.

[0036] In the embodiments of the present application, the words such as "example", "for example" and the like are used to represent as an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0037] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized. "Of", "corresponding" and "corresponding" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized.

[0038] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1, and the meanings expressed are consistent when the distinction is not emphasized.

[0039] In order to make the technical problems, technical solutions and advantages to be solved by the present application more clear, the following will be described in detail with reference to the drawings and specific embodiments.

[0040] As shown in Figure 1 The embodiments of the present application provide a key indicator extraction method for digital manufacturing maturity evaluation, which comprises the following steps: S1, obtaining intelligent factory workshop process business data for preprocessing, extracting a process business data candidate indicator dataset, performing preliminary dynamic screening analysis on the process business data candidate indicator dataset, eliminating the candidate indicator dataset of the business data, and recalibrating the candidate indicator dataset of the business data to obtain a first core candidate indicator set.

[0041] It should be noted that the extraction of each process business data candidate index dataset is as follows: in the preparation stage, the data flow from the material procurement system to the warehouse management system, automatic material taking and inventory synchronization, etc. are analyzed in depth, including the refining of business activities such as material demand planning automatic issuance rate, real-time confirmation rate of material taking instructions between ERP and MES systems, and inventory dynamic updating and automatic replenishment trigger frequency, and then the core business volume generated in these activities, such as material taking instruction response time, automatic replenishment execution accuracy and inventory data consistency ratio, are taken as the first core candidate index of the preparation stage; in the assembly stage, the task allocation of the assembly line, digital work order execution, and material arrival and assembly quality traceability are analyzed, and activities including assembly sequence automatic scheduling completion rate, process parameter automatic issuance and execution deviation compensation times, and instruction interaction success rate between MES and robot control system are identified, and the first core candidate index of the assembly stage such as assembly task automatic completion rate, work order and actual assembly data consistency rate, and assembly fault self-diagnosis coverage rate is refined from these activities; in the debugging stage, it is necessary to include the automatic debugging business activities of device and system self-checking, parameter optimization and on-site fault recovery, focusing on capturing the automatic parameter tuning script execution success rate, remote debugging command execution delay, and automatic identification and closing rate of debugging process alarms, and extracting business volume such as automatic parameter tuning coverage, alarm response one-time solution rate and debugging time compression ratio as the first core candidate index of the debugging stage; the key to the test stage is the running efficiency of the automatic test platform and the completeness of the test data closed loop, including iterative analysis of business activities such as automatic test case triggering rate, test self-execution script error rate and timeliness of automatically generated test reports, and extracting automatic test coverage, test result consistency rate, etc. as the first core candidate index of the test stage.

[0042] It should be noted that the production unit refers to a manufacturing unit that can be independently identified, monitored and evaluated in physical space or scheduling logic from a complete set of processes such as preparation, assembly, debugging and testing.

[0043] S2, analyze the first core candidate index set to obtain production unit evaluation factors of each core candidate index data in the first core candidate index set, screen to obtain business data first production unit evaluation level, and perform tolerance compensation on the business data.

[0044] S3, re-screen the business data after tolerance compensation to obtain the business data second production unit evaluation level grading result.

[0045] S4, according to the business data second production unit evaluation level grading result, adjust and analyze each process to generate candidate optimization schemes for each process.

[0046] Specifically, a preliminary dynamic screening analysis is performed on each process business data candidate indicator dataset, and the specific process is as follows: a sliding window is preset, each process business data candidate indicator dataset is extracted in the sliding window, and the completeness rate, extreme point quantity, median drift quantity, and drift mean value of the candidate indicator data of each process business data are obtained.

[0047] The completeness rate, extreme point quantity, median drift quantity, and drift mean value of the candidate indicator data of each process business data are compared with the defined completeness rate, defined extreme point quantity, defined median drift quantity, and defined drift mean value stored in the database, respectively, a proportion analysis is performed, and a weight coefficient is introduced to obtain a discrimination value of the business data candidate indicator dataset. The discrimination value of the business data candidate indicator dataset is used to quantitatively evaluate the effectiveness of the candidate indicator data of each process business data in distinguishing different manufacturing maturity stages.

[0048] It should be noted that the discrimination value of the business data candidate indicator dataset reflects the "sensitivity" and "representativeness" of the candidate indicator in multiple dimensions such as data quality (completeness rate) and dynamic characteristics (extreme point, drift quantity) of each process, thereby ensuring that the selected indicator can not only stably reflect the subtle differences in process operation state, but also has the ability to distinguish the overall maturity level, which helps the system to more objectively determine which candidate indicator is most suitable for marking and tracking the progress of the manufacturing system at each stage of the digital transformation path.

[0049] It should be noted that the discrimination value of the business data candidate indicator dataset is analyzed under the following conditions: , In the formula, Q1 represents the discrimination value of the business data candidate indicator dataset, W 1k represents the completeness rate of the kth process business data candidate indicator data, J 2k represents the extreme point quantity of the kth process business data candidate indicator data, Z 3k represents the median drift quantity of the kth process business data candidate indicator data, P 4k represents the drift mean value of the kth process business data candidate indicator data, W1 represents the defined completeness rate, J2 represents the defined extreme point quantity, Z3 represents the defined median drift quantity, P4 represents the defined drift mean value, A1 represents the weight coefficient corresponding to the completeness rate stored in the database, A2 represents the weight coefficient corresponding to the extreme point quantity stored in the database, A3 represents the weight coefficient corresponding to the median drift quantity stored in the database, A4 represents the weight coefficient corresponding to the drift mean value stored in the database, k represents the number of each process, k = 1, 2, 3,..., n, and n is the total number of processes.

[0050] It should be noted that the number of extreme points of the process business data candidate index data refers to the number of relative local maximum or minimum values in the data sequence, that is, the peak or valley points formed by the up and down of adjacent points, and not the points with zero derivative in the mathematical sense, which is used to reflect the activity degree of data fluctuation, and when the number of extreme points is less than 1, it is automatically set to 1 to avoid meaningless calculation due to zero extreme points, and to help ensure that the system can still output stable and comparable discrimination values even in the special scene of “no fluctuation”.

[0051] It should be noted that in the dynamic analysis of the candidate index, there is a complex mutual influence among the completeness rate, the number of extreme points, the median drift amount and the drift mean: when the completeness rate suddenly decreases, the missing data is often filled by interpolation or completion strategy, and also causes “drop” or “lift” effect at the boundary because the mean value is inconsistent with the real trend before and after, which is easy to be counted as an extreme value, thereby pushing up the number of extreme points; the frequent occurrence of extreme points repeatedly pulls the median in the sliding window, so that the median drift amount increases significantly; with the accumulation of the median drift amount, the drift mean in the sliding window also rises accordingly, reflecting the deviation of the overall trend and triggering the system to adjust the compensation threshold; once the drift mean maintains at a high level for a long time, the system will automatically tighten the data completion or filtering strategy to restore a higher completeness rate and re-remove the extreme points, so as to restore the completeness rate in the next round of evaluation; when the completeness rate is high and the number of extreme points falls to a controllable level, the median drift amount tends to be flat, and the drift mean falls accordingly, indicating that the data distribution has returned to internal stability, and then the completion rule and filtering bandwidth can be relaxed, allowing less human intervention.

[0052] It should be noted that the weight coefficient corresponding to the completeness rate, the weight coefficient corresponding to the number of extreme points, the weight coefficient corresponding to the median drift amount, and the weight coefficient corresponding to the drift mean are stored in the database and the value range is usually set between 0 and 1. For example, by constructing a mapping table between the completeness rate, the number of extreme points, the median drift amount, the drift mean and the weight coefficient respectively, the real-time detected completeness rate, the number of extreme points, the median drift amount and the drift mean are respectively input and transmitted to the corresponding mapping relationship table in the database, so as to quickly obtain the weight coefficient corresponding to the completeness rate, the weight coefficient corresponding to the number of extreme points, the weight coefficient corresponding to the median drift amount, and the weight coefficient corresponding to the drift mean.

[0053] Specifically, the candidate indicator data set of the business data is eliminated, and the specific process is as follows: the discrimination value of the candidate indicator data set of the business data is extracted, and is compared with the set discrimination threshold value of the candidate indicator data set of the business data; if the discrimination value of the candidate indicator data set of the business data is lower than the discrimination threshold value of the candidate indicator data set of the business data, the candidate indicator data set of the business data is eliminated. The discrimination deviation value of the candidate indicator data set of the business data after elimination of the candidate indicator of the business data is obtained by subtracting the discrimination threshold value of the candidate indicator data set of the business data from the discrimination value of the candidate indicator data set of the business data after elimination of the candidate indicator of the business data; if the discrimination deviation value of the candidate indicator data set of the business data after elimination of the candidate indicator of the business data is higher than the preset discrimination deviation threshold value of the candidate indicator data set of the business data, the elimination of the candidate indicator of the business data is abandoned, and the indicator is re-added to the core candidate set; if the discrimination deviation value of the candidate indicator data set of the business data after elimination of the candidate indicator of the business data is lower than or equal to the preset discrimination deviation threshold value of the candidate indicator data set of the business data, the candidate indicator of the business data is eliminated.

[0054] It should be noted that because some business data candidate indicators are eliminated, the discrimination value of the remaining indicator set will usually change, because the discrimination is based on the statistical characteristics and weight distribution of the entire indicator set. When a indicator is removed, it is equivalent to changing the normalization reference and weight structure, for example, the complete rate, the number of extreme points, the median drift and other statistical quantities may be adjusted, and the weight of each indicator in the total discrimination contribution will be redistributed, so that the elimination of a candidate indicator not only eliminates its own discrimination signal, but also indirectly changes other indicators, thereby making the new discrimination value or deviation value present different results.

[0055] Specifically, the candidate indicator data set of the business data is re-calibrated, and the specific process is as follows: the discrimination value of the candidate indicator data set of the business data is extracted, and is compared with the set discrimination threshold value of the candidate indicator data set of the business data; if the discrimination value of the candidate indicator data set of the business data is lower than the discrimination threshold value of the candidate indicator data set of the business data, the candidate indicator data set of the business data is re-calibrated, and the candidate indicator data set of the business data of each process is re-acquired, to obtain the first core candidate indicator set; if the discrimination value of the candidate indicator data set of the business data is higher than or equal to the discrimination threshold value of the candidate indicator data set of the business data, the candidate indicator data set of the business data does not need to be re-calibrated, and the first core candidate indicator set is directly obtained.

[0056] It should be noted that the recalibration of the candidate indicator data set of the business data is to clean up the dimensions with weak contribution to the maturity assessment and noise optimization after eliminating some candidate indicators of the business data, and to evaluate the discrimination of the candidate indicator set that has been filtered and summarized, and when the overall discrimination of the whole indicator set is still lower than the threshold, the sliding window, filtering and the like need to be adjusted to improve the overall discriminant sensitivity, otherwise it means that the whole set has sufficient discrimination ability and does not need to be calibrated.

[0057] It should be noted that the recalibration of the candidate indicator data set of the business data is to clean up the dimensions with weak contribution to the maturity assessment and noise optimization after eliminating some candidate indicators of the business data, and to evaluate the discrimination of the candidate indicator set that has been filtered and summarized, and when the overall discrimination of the whole indicator set is still lower than the threshold, the sliding window, filtering and the like need to be adjusted to improve the overall discriminant sensitivity, otherwise it means that the whole set has sufficient discrimination ability and does not need to be calibrated.

[0058] Specifically, the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set is obtained, and the specific process is as follows: obtaining the value of each core candidate indicator data in the first core candidate indicator set, and extracting the threshold value corresponding to each core candidate indicator data value stored in the database.

[0059] The average Pearson correlation coefficient of each core candidate indicator data in the first core candidate indicator set and the target business volume stored in the database is obtained, denoted as the average Pearson correlation coefficient of each core candidate indicator data, and the Pearson correlation coefficient threshold is extracted.

[0060] The data values of each core candidate indicator in the first core candidate indicator set are compared with the threshold value corresponding to each core candidate indicator data value, the average Pearson correlation coefficient of each core candidate indicator data and the Pearson correlation coefficient threshold value corresponding thereto, and the discrimination value and weight coefficient of the business data candidate indicator data set are introduced to obtain the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set. The production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set is used to quantitatively evaluate the performance intensity of the production unit in each key dimension of digital manufacturing maturity.

[0061] It should be noted that evaluating the performance intensity of the production unit in each key dimension of digital manufacturing maturity is specifically to measure the comprehensive performance intensity of the production unit in each key dimension of digital manufacturing maturity (such as data quality, process stability, automation linkage and intelligent feedback, etc.), to obtain the maturity level of the whole production unit, and to provide accurate direction for subsequent targeted improvement.

[0062] It should be noted that the process of obtaining the average Pearson correlation coefficient of each core candidate indicator data in the first core candidate indicator set and the target business volume stored in the database is as follows: align the candidate indicators and the original time series data of the target business volume (such as production efficiency, defect rate, delivery timeliness) according to the unified time stamp or through the dynamic time warping (DTW) method, if necessary, interpolate or downsample the data of different sampling frequencies to ensure consistent time steps; then select the sliding window length (for example, 8 hours of a shift or 500 sampling records) and the step (such as every 1 hour or 100 records) and start from the first time point, continuously slide the entire time series with the step, extract the subsequence of the candidate indicator and the corresponding business target in each window; for each window subsequence, calculate the linear correlation strength of the two using the Pearson correlation coefficient formula to obtain a set of window correlation values, and after summarizing and counting these windowed correlation values, the average Pearson correlation coefficient of the entire indicator is obtained. The entire Pearson time-varying correlation analysis based on sliding window can not only quantify the global relationship between the indicator and the business target, but also seamlessly capture the dynamic ability of the indicator in the production process, providing solid data support for subsequent indicator selection, elimination or reconstruction.

[0063] It should be noted that the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set is specifically analyzed under the following conditions: , In the formula, Q 2j represents the production unit evaluation factor of the jth core candidate indicator data in the first core candidate indicator set, B 1j represents the data value of the jth core candidate indicator in the first core candidate indicator set, r2j an average Pearson correlation coefficient of the jth core candidate indicator data, a discrimination value of the business data candidate indicator data set, B 2j a threshold value corresponding to the jth core candidate indicator data value, r j an average Pearson correlation coefficient of the jth core candidate indicator data, C1 represents a weight coefficient corresponding to the core candidate indicator data value stored in the database, C2 represents a weight coefficient corresponding to the Pearson correlation coefficient stored in the database, C3 represents a weight coefficient corresponding to the discrimination value of the business data candidate indicator data set stored in the database, j represents the number of each core candidate indicator data, j = 1, 2, 3, …, m, and m is the total number of core candidate indicator data.

[0064] It should be noted that in the construction of the production unit evaluation factor, there is also a complex interaction among the real-time data value of each core candidate indicator, its average Pearson correlation coefficient with the business target, and the previously calculated discrimination value of the indicator: when the real-time value of a certain indicator deviates significantly (for example, the assembly automation completion rate suddenly drops), if the fluctuation is accompanied by an increase in the correlation coefficient with production efficiency or defect rate, it means that the fluctuation is "in resonance" with the business results, and at this time the high discrimination value will amplify the contribution of the indicator in the evaluation factor, making the evaluation result extremely sensitive to the fluctuation; on the contrary, if the indicator value deviates but the correlation coefficient decreases, it means that the fluctuation is more likely to be caused by noise or an occasional event, and the discrimination value will weaken the weight of the indicator in the evaluation, thereby avoiding excessive fluctuations in the evaluation factor; at the same time, when the discrimination value itself is at a high level and both the indicator value and the correlation coefficient are changing, the evaluation factor will appear a short-term transition, prompting the system to pay close attention to the indicator; if all three return in the same direction, it proves that the adjustment or optimization measures have been effective, and the evaluation factor will quickly return to normal; through this synergistic dynamic response mechanism, the system can ensure high sensitivity to major business fluctuations while suppressing false positives caused by noise or single-dimensional fluctuations, thereby achieving precise control of the digital maturity of the production unit.

[0065] Specifically, the business data first production unit evaluation level is obtained by screening, and the specific process is as follows: extracting the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set, and comparing it with the production unit evaluation factor interval stored in the database, counting the number of production unit evaluation factors of each core candidate indicator data in the first core candidate indicator set that exceed the production unit evaluation factor interval, and recording it as the number of out-of-limit indicators.

[0066] It should be noted that the closer the production unit evaluation factor of each core candidate index data in the first core candidate index set to the center value in the interval, the higher the reference value of the maturity evaluation. If it deviates from the interval in any direction, it may mean that the index has too strong volatility, correlation failure or abnormal behavior, thereby affecting the accurate determination of the overall maturity of the production unit.

[0067] The number of over-limit indicators is extracted and compared with the unit evaluation level corresponding to each interval of the set number of over-limit indicators to obtain the business data first production unit evaluation level.

[0068] The business data first production unit evaluation level includes first, second and third levels.

[0069] It should be noted that the first, second and third levels correspond to different levels of the business data first production unit in the digital manufacturing maturity evaluation, wherein the first level indicates that the production unit has stable performance of each core candidate index, the evaluation factor is basically within the reasonable interval, and has higher digital capability and maturity; the second level indicates that there are some indicators exceeding the limit but the overall still maintains in the controllable range, indicating that the production unit has certain digital support capability but needs local optimization; the third level indicates that multiple core indicators evaluation factors exceed the reasonable interval, reflecting that the production unit has obvious short board in digital monitoring, response, coordination and other aspects, and needs systematic improvement or strategic adjustment.

[0070] Specifically, the business data is compensated for tolerance, and the specific process is: extracting the absolute value of the difference between the production unit evaluation factor of each core candidate index data in the first core candidate index set and the nearest boundary of the production unit evaluation factor interval, and compensating the business data for tolerance according to the absolute value of the difference between the production unit evaluation factor of each core candidate index data in the first core candidate index set and the nearest boundary of the production unit evaluation factor interval.

[0071] It should be noted that the specific process of tolerance compensation for business data is as follows: the absolute value of the difference between the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set and the nearest boundary of the production unit evaluation factor interval is recorded as the deviation value of each core candidate indicator data in the first core candidate indicator set. First, it is judged whether the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set falls within the range of the evaluation factor interval stored in the database. If it is within the interval, no tolerance compensation is performed; if it is outside the interval, the difference between the evaluation factor and the nearest boundary of the interval is extracted, and a hierarchical tolerance compensation strategy is implemented according to the range of the absolute value of the difference: if the absolute value of the difference is within 5%, and the evaluation factor is higher than the upper limit boundary, then the sliding window length is moderately increased by 3% to 5%, and the filter bandwidth is increased by 0.5% to 1%, to smooth the abnormal peak caused by sudden rise; if the evaluation factor is lower than the lower limit boundary, the sliding window length can still be increased by 3% to 5%, but the increase of the filter bandwidth can be controlled between 0.3% and 0.8% to buffer the fluctuations caused by abnormal decrease; if the absolute value of the difference is between 5% and 10%, and the evaluation factor is lower than the lower limit boundary of the corresponding evaluation interval, then the sliding window length is increased by 5% to 10%, and the filter bandwidth is increased by about 1% to 2% simultaneously, to improve the coverage ability of the system to short-term abnormalities; if the evaluation factor is higher than the upper limit boundary of the corresponding evaluation interval, then the sliding window length is increased by 5% to 10%, but the filter bandwidth is increased by about 2% to 3%, which focuses more on suppressing the disturbance of frequent oscillation on the evaluation result; if the absolute value of the difference exceeds 10%, then the sliding window length is further increased by about 10% to 20%; wherein, if the evaluation factor is lower than the lower limit boundary of the interval, then the filter bandwidth is increased by about 2% to 3% to enhance the smoothing ability of occasional abnormalities in the downward trend; if the evaluation factor is higher than the upper limit boundary of the interval, then the filter bandwidth is increased by about 3% to 4% to suppress the misjudgment caused by sudden peak. This compensation mechanism ensures that when the evaluation factor exceeds the interval, the system can implement targeted parameter adaptive correction according to its deviation degree and direction, avoiding misjudgment and enhancing robustness.

[0072] It should be noted that these adjustment intervals (such as 3%-5%, 0.5%-1%, 5%-10%, etc.) are based on statistical analysis and engineering experience of a large amount of business data in historical real scenarios, and the essence is to ensure data authenticity while avoiding misjudgment or false positives due to short-term abnormalities; the smaller the deviation, the less adjustment is needed to prevent over-response, and the greater the deviation, the more significant the fluctuation, and the compensation range of the sliding window and filter should also be gradually increased; these adjustment ranges can be adapted and optimized according to the sensitivity of different enterprises or processes; usually in practical applications, the median or empirical upper and lower limits of each interval are used as the initial setting, for example, "3%-5%" can be set to 4% as the initial adjustment value, and in actual operation, it can also be fine-tuned in combination with historical compensation effects.

[0073] Specifically, the tolerance-compensated business data is reclassified to obtain a business data second production unit evaluation level classification result, and the specific process is as follows: the number of out-of-limit indicators corresponding to the tolerance-compensated business data is counted, and if the number of out-of-limit indicators corresponding to the tolerance-compensated business data is higher than or equal to a set threshold of the number of out-of-limit indicators corresponding to the tolerance-compensated business data, the business data first production unit evaluation level is downgraded, thereby obtaining the business data second production unit evaluation level classification result.

[0074] If the number of out-of-limit indicators corresponding to the tolerance-compensated business data is lower than the set threshold of the number of out-of-limit indicators corresponding to the tolerance-compensated business data, and the business data first production unit evaluation level is level one, the business data first production unit evaluation level is maintained; if the number of out-of-limit indicators corresponding to the tolerance-compensated business data is lower than the set threshold of the number of out-of-limit indicators corresponding to the tolerance-compensated business data, and the business data first production unit evaluation level is not level one, the business data first production unit evaluation level is upgraded, thereby obtaining the business data second production unit evaluation level classification result.

[0075] Specifically, according to the business data second production unit evaluation level classification result, each process is analyzed to generate a candidate optimization scheme for each process, and the specific process is as follows: the business data second production unit evaluation level classification result is extracted, and if the business data second production unit evaluation level classification result is not level one, a simulation model trained in advance in the digital twin environment is called, process parameters are input to a virtual production line, and a candidate optimization scheme for each process is generated.

[0076] It should be noted that the simulation model trained in the digital twin environment is a multivariate prediction and inversion model constructed on the basis of real historical production data, equipment operation data and process parameter data, combined with multi-dimensional performance indicators (such as capacity, yield, energy consumption, equipment utilization rate, etc.). Through long-term dynamic iterative training in the digital twin platform, the model has strong sensitivity to various process changes and expected output simulation capabilities. The key process parameters of the current production unit (such as temperature set value, assembly rhythm, test threshold, electrical drive setting, logistics interval time, etc.), the abnormal index and its deviation value corresponding to the second production unit evaluation level, the current production load level, the inter-process coordination index (such as automatic line synchronization rate or robot scheduling delay) are input into the simulation model. Through these inputs, the running dynamics and index response trend of each process under the current process configuration can be accurately simulated; the output of the model is a set of candidate optimization schemes corresponding to each process.

[0077] It should be noted that if the second production unit evaluation level grading result is level one, the digital twin simulation model does not need to be called immediately for process intervention. At this time, the system will enter the monitoring and maintenance stage.

[0078] It should be noted that the generated group candidate optimization schemes include increasing the number of robot assembly grasps, shortening the logical branches of the debugging script execution, improving the parallelism of the test platform, or adjusting the scanning frequency of the factory inspection; for each optimization scheme, the digital twin model will output the corresponding key candidate index change prediction and the comprehensive impact evaluation on overall production efficiency, defect rate and delivery timeliness within seconds, and these prediction results will be sorted by priority and fed back to the field.

[0079] As shown in Figure 2 , the system will enter the monitoring and maintenance stage. Figure 2As shown in the schematic diagram of the logical flow of the key indicator extraction method for digital manufacturing maturity evaluation, first, the business data of each process is obtained and preprocessed, and then the initial candidate indicator dataset is extracted; then through dynamic screening, indicators with strong instability or noise sensitivity are removed. Next, it is judged whether the remaining indicators meet the preset discrimination threshold, if yes, the indicators are calibrated to form the first core candidate indicator set; if not, the current indicators are directly used to form the first core candidate indicator set. Based on the first core candidate indicator set, the production unit evaluation factor of each indicator is obtained, and the first stage maturity level is obtained accordingly. Then it is judged whether there are out-of-limit indicators whose evaluation factors exceed the preset interval, if yes, the tolerance compensation mechanism is triggered, the analysis window and the filter setting are adjusted to smooth the abnormal influence, the data is screened again to obtain the second stage evaluation level. Finally, the system calls the simulation model according to the results of the second evaluation level to generate the corresponding process-level optimization scheme. Through the construction of a complete closed-loop process from business data collection, candidate indicator screening to production unit maturity level evaluation, the dynamic extraction and accurate calibration of the key indicators of digital manufacturing maturity are realized. The process innovatively introduces an indicator discrimination judgment mechanism, constructs a discrimination value based on data integrity rate, number of extreme points and median drift, and sets a reasonable threshold to remove redundant or low-value indicators, ensuring that each core indicator remaining has actual contribution to distinguishing digital stages. At the same time, through the evaluation factor construction mechanism, each core indicator is related to the macro target (such as efficiency, quality), making the evaluation level more representative of the business. Through the tolerance compensation link, dynamic window adjustment and filter bandwidth adjustment technology are introduced, which ensures the sensitivity of evaluation while reducing the misjudgment caused by short-term fluctuations. Finally, the system generates targeted process optimization suggestions based on the second evaluation level results, significantly improving the scientificity, adaptability and evaluation reliability of the key indicator extraction of digital manufacturing, providing strong quantitative support and decision basis for enterprises to dynamically grasp the digital evolution state.

[0080] As shown in the schematic diagram of the logical flow of the key indicator extraction method for digital manufacturing maturity evaluation, first, the business data of each process is obtained and preprocessed, and then the initial candidate indicator dataset is extracted; then through dynamic screening, indicators with strong instability or noise sensitivity are removed. Next, it is judged whether the remaining indicators meet the preset discrimination threshold, if yes, the indicators are calibrated to form the first core candidate indicator set; if not, the current indicators are directly used to form the first core candidate indicator set. Based on the first core candidate indicator set, the production unit evaluation factor of each indicator is obtained, and the first stage maturity level is obtained accordingly. Then it is judged whether there are out-of-limit indicators whose evaluation factors exceed the preset interval, if yes, the tolerance compensation mechanism is triggered, the analysis window and the filter setting are adjusted to smooth the abnormal influence, the data is screened again to obtain the second stage evaluation level. Finally, the system calls the simulation model according to the results of the second evaluation level to generate the corresponding process-level optimization scheme. Through the construction of a complete closed-loop process from business data collection, candidate indicator screening to production unit maturity level evaluation, the dynamic extraction and accurate calibration of the key indicators of digital manufacturing maturity are realized. The process innovatively introduces an indicator discrimination judgment mechanism, constructs a discrimination value based on data integrity rate, number of extreme points and median drift, and sets a reasonable threshold to remove redundant or low-value indicators, ensuring that each core indicator remaining has actual contribution to distinguishing digital stages. At the same time, through the evaluation factor construction mechanism, each core indicator is related to the macro target (such as efficiency, quality), making the evaluation level more representative of the business. Through the tolerance compensation link, dynamic window adjustment and filter bandwidth adjustment technology are introduced, which ensures the sensitivity of evaluation while reducing the misjudgment caused by short-term fluctuations. Finally, the system generates targeted process optimization suggestions based on the second evaluation level results, significantly improving the scientificity, adaptability and evaluation reliability of the key indicator extraction of digital manufacturing, providing strong quantitative support and decision basis for enterprises to dynamically grasp the digital evolution state. Figure 3

[0081] The tolerance compensation module is used to analyze the first core candidate indicator set, obtain the production unit evaluation factor of each core candidate indicator data in the first core candidate indicator set, screen to obtain the first production unit evaluation level of the business data, and perform tolerance compensation on the business data.​

[0082] A production unit level confirmation module is configured to re-screen the tolerance-compensated service data to obtain a service data second production unit evaluation level classification result.

[0083] An optimization procedure adjustment module is configured to analyze and adjust each procedure according to the service data second production unit evaluation level classification result to generate a candidate optimization scheme for each procedure.

[0084] It should also be appreciated that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically EPROM (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM) used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM) and direct rambus RAM (DR RAM).

[0085] The above-described embodiments can be implemented in whole or in part by software, hardware (such as a circuit), firmware, or any combination thereof. When implemented in software, the above-described embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions according to the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center through a wired (for example, infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0086] It should be understood that the term "and / or" herein merely describes an association relationship of associated objects, which means that there can be three relationships, for example, A and / or B can represent three cases of A alone, A and B together, and B alone, where A and B can be singular or plural. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after it, but it can also represent an "and / or" relationship, which can be understood in the context before and after it.

[0087] In the present application, "at least one" means one or more, and "multiple" means two or more. "At least one of the following" or the like means any combination of the items, including any combination of single item or multiple items. For example, at least one of a, b, or c can represent a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, and c can be single or multiple.

[0088] It should be understood that in various embodiments of the present application, the size of the sequence number of the above-described processes does not mean the order of execution, and the execution order of the processes should be determined by their functions and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0089] Those skilled in the art can clearly understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0090] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the devices, apparatuses and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0091] In several embodiments provided by the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic, for example, the division of units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0092] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0093] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically independently, or two or more units can be integrated into one unit.

[0094] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the present application that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0095] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for extracting key indicators for digital manufacturing maturity evaluation, characterized in that: The method comprises: S1, obtaining business data of each process in the smart factory workshop and preprocessing it, extracting candidate indicator data sets of each process business data, performing preliminary dynamic screening and analysis on the candidate indicator data sets of each process business data, eliminating the candidate indicator data sets of the business data, and recalibrating the candidate indicator data sets of the business data to obtain a first core candidate indicator set; S2, analyzing the first core candidate indicator set, obtaining the production unit assessment factor of each core candidate indicator data in the first core candidate indicator set, screening to obtain the first production unit assessment level of the business data, and performing tolerance compensation on the business data; S3, re-screening the business data after tolerance compensation to obtain the second production unit assessment grade grading result of the business data; S4, adjusting and analyzing each process based on the second production unit evaluation and grading results of the business data, and generating candidate optimization plans for each process.

2. The key indicator extraction method for digital manufacturing maturity evaluation according to claim 1 is characterized in that: The preliminary dynamic screening and analysis of the candidate indicator data sets of each process business data is carried out as follows: A sliding window is preset, and candidate indicator data sets of each process business data are extracted in the sliding window, and the completeness rate, number of extreme points, median drift and drift mean of the candidate indicator data of each process business data are obtained; The completeness rate of the candidate indicator data of the business data of each process is extracted and compared with the defined completeness rate stored in the database, the number of extreme points and the number of defined extreme points stored in the database, the median drift and the defined median drift stored in the database, and the drift mean and the defined drift mean stored in the database, and a proportion analysis is performed respectively and a weight coefficient is introduced to obtain the discrimination value of the business data candidate indicator data set. The discrimination value of the business data candidate indicator data set is used to quantitatively evaluate the effectiveness of extracting the candidate indicator data of the business data of each process in distinguishing different manufacturing maturity stages.

3. The key indicator extraction method for digital manufacturing maturity evaluation according to claim 1 is characterized in that: The specific process of eliminating candidate indicator data sets for business data is as follows: Extract the discrimination value of the candidate indicator dataset of the business data, and compare it with the discrimination threshold of the candidate indicator dataset of the business data. If the discrimination value of the candidate indicator dataset of the business data is lower than the discrimination threshold of the candidate indicator dataset of the business data, the candidate indicator dataset of the business data is eliminated; A candidate indicator data detection cycle is preset, and the discrimination value of the business data candidate indicator data set corresponding to each business data after each business data candidate indicator is eliminated is obtained during the candidate indicator data detection cycle. The discrimination value of the business data candidate indicator data set corresponding to each business data after the business data candidate indicators are eliminated is subtracted from the discrimination threshold of the business data candidate indicator data set to obtain the discrimination deviation value of the business data candidate indicator data set corresponding to each business data after the business data candidate indicators are eliminated. If the discrimination deviation value of the business data candidate indicator data set corresponding to the business data candidate indicators of a certain business data after the business data candidate indicators are eliminated is higher than the preset discrimination deviation threshold of the business data candidate indicator data set, the business data candidate indicator is abandoned and the indicator is added back to the core candidate set. If the discrimination deviation value of the business data candidate indicator data set corresponding to the business data candidate indicators of a certain business data after the business data candidate indicators are eliminated is lower than or equal to the preset discrimination deviation threshold of the business data candidate indicator data set, the business data candidate indicator is eliminated.

4. The key indicator extraction method for digital manufacturing maturity evaluation according to claim 1 is characterized in that: The specific process of recalibrating the candidate indicator dataset of the business data is as follows: The discrimination value of the business data candidate indicator dataset is extracted and compared with the set discrimination threshold of the business data candidate indicator dataset. If the discrimination value of the business data candidate indicator dataset is lower than the discrimination threshold of the business data candidate indicator dataset, the candidate indicator dataset of the business data is recalibrated and the business data candidate indicator dataset of each process is re-acquired to obtain the first core candidate indicator set. If the discrimination value of the business data candidate indicator dataset is higher than or equal to the discrimination threshold of the business data candidate indicator dataset, the candidate indicator dataset of the business data does not need to be recalibrated and the first core candidate indicator set is directly obtained.

5. The method for extracting key indicators for digital manufacturing maturity evaluation according to claim 1 is characterized in that: The specific process of obtaining the production unit assessment factor of each core candidate indicator data in the first core candidate indicator set is as follows: Obtaining data values ​​of each core candidate indicator in the first core candidate indicator set, and extracting threshold values ​​corresponding to each core candidate indicator data value stored in a database; Obtaining an average Pearson correlation coefficient between each core candidate indicator data in the first core candidate indicator set and the target business volume stored in the database, recording it as the average Pearson correlation coefficient of each core candidate indicator data, and extracting a Pearson correlation coefficient threshold; The data values ​​of each core candidate indicator in the first core candidate indicator set are compared with the thresholds corresponding to the data values ​​of each core candidate indicator, and the average Pearson correlation coefficient of each core candidate indicator data and its corresponding Pearson correlation coefficient threshold, and the discrimination value and weight coefficient of the business data candidate indicator data set are introduced to obtain the production unit assessment factor of each core candidate indicator data in the first core candidate indicator set. The production unit assessment factor of each core candidate indicator data in the first core candidate indicator set is used to quantitatively evaluate the performance strength of the production unit in each key dimension of digital manufacturing maturity.

6. The method for extracting key indicators for digital manufacturing maturity evaluation according to claim 1, characterized in that: The screening process to obtain the first production unit assessment level of the business data is as follows: Extracting the production unit assessment factor of each core candidate indicator data in the first core candidate indicator set, and comparing it with the production unit assessment factor interval stored in the database, counting the number of production unit assessment factors of each core candidate indicator data in the first core candidate indicator set that exceed the production unit assessment factor interval, and recording it as the number of exceeded indicators; Extract the number of over-limit indicators and compare it with the unit assessment level corresponding to each interval of the set over-limit indicator number to obtain the first production unit assessment level of the business data; The business data first production unit assessment levels include level one, level two, and level three.

7. The method for extracting key indicators for digital manufacturing maturity evaluation according to claim 6, characterized in that: The specific process of performing tolerance compensation on business data is as follows: The absolute value of the difference between the production unit assessment factor of each core candidate indicator data in the first core candidate indicator set and the nearest boundary of the production unit assessment factor interval is extracted, and tolerance compensation is performed on the business data based on the absolute value of the difference between the production unit assessment factor of each core candidate indicator data in the first core candidate indicator set and the nearest boundary of the production unit assessment factor interval.

8. The method for extracting key indicators for digital manufacturing maturity evaluation according to claim 1, characterized in that: The business data after tolerance compensation is re-screened to obtain the second production unit assessment grade grading result of the business data. The specific process is as follows: Counting the number of over-limit indicators corresponding to the business data after tolerance compensation; if the number of over-limit indicators corresponding to the business data after tolerance compensation is higher than or equal to the set threshold number of over-limit indicators corresponding to the business data after tolerance compensation, the assessment grade of the first production unit of the business data is lowered, thereby obtaining the assessment grade result of the second production unit of the business data; If the number of over-limit indicators corresponding to the business data after tolerance compensation is lower than the set threshold value of the number of over-limit indicators corresponding to the business data after tolerance compensation, and the assessment level of the first production unit of the business data is level one, the assessment level of the first production unit of the business data is maintained; if the number of over-limit indicators corresponding to the business data after tolerance compensation is lower than the set threshold value of the number of over-limit indicators corresponding to the business data after tolerance compensation, and the assessment level of the first production unit of the business data is not level one, the assessment level of the first production unit of the business data is adjusted upward, thereby obtaining the assessment level grading result of the second production unit of the business data.

9. The method for extracting key indicators for digital manufacturing maturity evaluation according to claim 1, characterized in that: The process of adjusting and analyzing each process based on the second production unit evaluation results of the business data to generate candidate optimization plans for each process is as follows: Extract the grading results of the second production unit of business data. If the grading results of the second production unit of business data are not level one, call the simulation model pre-trained in the digital twin environment, input the process parameters into the virtual production line, and generate candidate optimization plans for each process.

10. A system using the key indicator extraction method for digital manufacturing maturity evaluation according to any one of claims 1 to 9, characterized in that: include: The candidate indicator collection and calibration module is used to obtain the business data of each process in the smart factory workshop for preprocessing, extract the candidate indicator data set of each process business data, perform preliminary dynamic screening and analysis on the candidate indicator data set of each process business data, eliminate the candidate indicator data set of the business data, and recalibrate the candidate indicator data set of the business data to obtain the first core candidate indicator set; A tolerance compensation module is used to analyze the first core candidate indicator set, obtain the production unit assessment factor of each core candidate indicator data in the first core candidate indicator set, screen to obtain the first production unit assessment level of the business data, and perform tolerance compensation on the business data; The production unit grade confirmation module is used to re-screen the business data after tolerance compensation to obtain the second production unit assessment grade grading result of the business data; The optimization process adjustment module is used to adjust and analyze each process based on the second production unit assessment grade grading results of business data and generate candidate optimization plans for each process.

Images