A feed whole-process production data acquisition method and system for a smart factory

Abstract

The present application relates to the technical field of data processing, in particular to a feed whole-process production data acquisition method and system for a smart factory, comprising: obtaining production data corresponding to each data acquisition device in each production stage in a feed production process; obtaining a position influence coefficient of each data acquisition device in each production stage according to data differences between production data corresponding to adjacent data acquisition devices in each production stage, in combination with position information of the data acquisition devices; obtaining a data condition effective index of each production stage according to similarity between position influence coefficients of different data acquisition devices in each production stage, in combination with fluctuation of environmental information; and analyzing whether to adjust a data acquisition frequency according to the data condition effective index of each production stage, to determine a data acquisition method. The present application can adaptively adjust the data acquisition frequency to optimize the data acquisition method, and improve the effectiveness of data acquisition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a feed whole-process production data acquisition method and system of a smart factory. BACKGROUND

[0002] With the development of livestock breeding industry, complete feed production line has become one of the essential tools to improve breeding efficiency. The working process of complete feed production line mainly includes five key processes: cleaning, crushing, mixing, granulation and packaging, forming a complete and efficient production system. In the whole-process production of feed in a smart factory, Internet of Things technology and data acquisition system play an important role. Internet of Things technology connects various sensors and devices, monitors key parameters such as temperature, humidity, flow and pressure, enables real-time data transmission and monitoring, ensures the normal operation of production environment and equipment, and realizes remote management and control.

[0003] In the actual multi-type feed production process, different proportions will cause some physical property changes of multi-link feed, including water content, granularity and other types, thereby affecting the feed raw material operation efficiency and quality in this production link. The specific influence mode is adhesion, accumulation, impact and other conditions of the conveying pipeline, which leads to reduced conveying efficiency, decreased granulation quality, and even intensified equipment vibration, which not only affects the feed production efficiency and quality, but also affects the service life of the equipment. At this time, the fixed frequency data acquisition process in the existing method is easily affected by objective factors such as environment and position, so that the effectiveness of data acquisition is poor. SUMMARY

[0004] In order to solve the technical problem of poor effectiveness of feed whole-process production data acquisition result in the existing method, the purpose of the present application is to provide a feed whole-process production data acquisition method and system of a smart factory, and the technical solution adopted is as follows:

[0005] In the first aspect, the present application provides a feed whole-process production data acquisition method of a smart factory, comprising:

[0006] acquiring production data corresponding to each data acquisition device in each production stage in the feed production process through different data acquisition devices in each production stage;

[0007] obtaining a position influence coefficient of each data acquisition device in each production stage according to the difference of data monitoring between the production data corresponding to the data acquisition devices at adjacent positions in each production stage, and combining the position information of each data acquisition device in each production stage;

[0008] According to the similarity between the position influence coefficients of different data collection devices in each production stage, and in combination with the fluctuation of environmental information in each production stage, an effective data status index of each production stage is obtained.

[0009] According to the data status effective index of each production stage, whether to adjust the data collection frequency is analyzed, and the data collection method of each data collection device is determined.

[0010] Preferably, the difference of data monitoring between the production data corresponding to the data collection devices at adjacent positions in each production stage is combined with the position information of each data collection device in each production stage to obtain the position influence coefficient of each data collection device in each production stage, specifically including:

[0011] According to the difference between the fluctuation of the production data corresponding to each data collection device in each production stage and the fluctuation of the production data corresponding to the data collection devices at adjacent positions, a position correlation factor of each data collection device in each production stage is obtained.

[0012] Based on the distance between the position of each data collection device in each production stage and the output position of the corresponding production stage, a relative distance factor of each data collection device in each production stage is determined.

[0013] Based on the ratio between the position correlation factor and the relative distance factor, the position influence coefficient of each data collection device in each production stage is determined.

[0014] Preferably, the difference between the fluctuation of the production data corresponding to each data collection device in each production stage and the fluctuation of the production data corresponding to the data collection devices at adjacent positions is used to obtain the position correlation factor of each data collection device in each production stage, specifically including:

[0015] For each production stage, based on the difference between each production data corresponding to each data collection device and the mean value of all production data, the monitoring difference degree of each production data corresponding to each data collection device is determined.

[0016] Based on the difference between the mean value of the monitoring difference degrees of all production data corresponding to each data collection device and the mean value of the monitoring difference degrees of all production data corresponding to the data collection devices at adjacent positions, the position correlation factor of each data collection device in each production stage is determined.

[0017] Preferably, the similarity between the position influence coefficients of different data collection devices in each production stage is combined with the fluctuation of environmental information in each production stage to obtain an effective data status index of each production stage, specifically including:

[0018] Based on the similarity between the position influence coefficients of different data acquisition devices in each production stage, the data acquisition devices in the same production stage are clustered, and the category fluctuation coefficient of each production stage is obtained according to the aggregation of the data acquisition devices in the corresponding clustering result of each production stage;

[0019] Based on the position influence coefficient and the category fluctuation coefficient of all data acquisition devices in each production stage, the abnormal sensitivity factor of each production stage is determined;

[0020] The environmental data of the position of each data acquisition device in each production stage under different production parameters is obtained to form the environmental data sequence of each production stage under different production parameters;

[0021] According to the difference between the environmental data sequence of each production stage under different production parameters and the balanced situation of the environmental data sequence of the same production stage under all production parameters, and combined with the abnormal sensitivity factor, the data status effective index of each production stage is obtained.

[0022] Preferably, the category fluctuation coefficient of each production stage is obtained according to the aggregation of the data acquisition devices in the corresponding clustering result of each production stage, specifically including:

[0023] For each production stage, the proportion of the number of data acquisition devices contained in each clustering cluster in the clustering result is obtained as the category proportion of each clustering cluster in the clustering result, and the discrete degree of all category proportions in the production stage is taken as the category fluctuation coefficient of the corresponding production stage.

[0024] Preferably, the data status effective index of each production stage is obtained according to the difference between the environmental data sequence of each production stage under different production parameters and the balanced situation of the environmental data sequence of the same production stage under all production parameters, and combined with the abnormal sensitivity factor, specifically including:

[0025] Based on the mean value of the environmental data of the same data acquisition device in each production stage under all production parameters, the environmental balanced sequence is determined, and based on the difference distance between the environmental data sequence of each production stage and the environmental balanced sequence under all production parameters and the ratio relationship between the abnormal sensitivity factor, the data status effective index of each production stage is determined.

[0026] Preferably, the abnormal sensitivity factor of each production stage is determined based on the position influence coefficient and the category fluctuation coefficient of all data acquisition devices in each production stage, specifically including:

[0027] The ratio between the mean value of the position influence coefficient of all data acquisition devices in each production stage and the category fluctuation coefficient is taken as the abnormal sensitivity factor of each production stage.

[0028] Preferably, the data condition effective index of each production stage is used to analyze whether to adjust the data collection frequency, and the data collection method of each data collection device is determined, specifically including:

[0029] When the data condition effective index of each production stage does not meet the effectiveness condition, the data collection operation is performed after adjusting the data collection frequency of the data collection device of the production stage; when the data condition effective index of each production stage meets the effectiveness condition, the data collection operation is performed at the current data collection frequency.

[0030] Preferably, the effectiveness condition is specifically that the data condition effective index of the production stage is greater than or equal to a preset effective threshold.

[0031] In a second aspect, the present application provides a feed whole-process production data collection system of a smart factory, which is used to realize the steps of a feed whole-process production data collection method of a smart factory, and specifically includes:

[0032] A data collection module is used to obtain production data corresponding to each data collection device through different data collection devices in each production stage of the feed production process.

[0033] A position feature analysis module is used to obtain a position influence coefficient of each data collection device in each production stage according to the difference between the data monitoring of the production data corresponding to the adjacent data collection devices in each production stage, and the position information of each data collection device in each production stage.

[0034] An effective feature analysis module is used to obtain a data condition effective index of each production stage according to the similarity between the position influence coefficients of different data collection devices in each production stage, and the fluctuation of the environmental information in each production stage.

[0035] A collection method determination module is used to analyze whether to adjust the data collection frequency according to the data condition effective index of each production stage, and determine the data collection method of each data collection device.

[0036] The embodiments of the present application have at least the following beneficial effects:

[0037] The application firstly acquires production data of each data acquisition device in different production stages, and provides a data basis for subsequent adaptive feature analysis process for different production stages. Then, the data monitoring difference between adjacent data acquisition positions in the same production stage is analyzed, and the position influence coefficient of the data acquisition device is preliminarily determined in combination with the corresponding data acquisition position information, so as to measure the direct influence of the data acquisition position corresponding to the data abnormality on the output quality. Further, on the basis of analyzing the similarity of the position influence degree corresponding to different data acquisition devices in the same production stage, the data effective degree of real-time monitoring in each production stage is comprehensively reflected by further combining the change of the environmental information, and finally the data condition effective index can be used to analyze whether the data acquisition frequency needs to be adjusted to optimize the data acquisition method. The application can adaptively adjust the data acquisition frequency to optimize the data acquisition method by deeply analyzing the position information change relationship of the data acquisition device, improve the effectiveness of data acquisition to a certain extent, and enable the production enterprise to realize real-time monitoring and optimization of the proportion. The data monitoring process can ensure that the physical properties of the feed meet the standards and improve the production quality of the feed. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without any creative effort.

[0039] Figure 1 is a step flow chart of a feed whole-process production data acquisition method of a smart factory provided by the present application;

[0040] Figure 2 is a step flow chart of a position influence coefficient acquisition method provided by the present application;

[0041] Figure 3 is a step flow chart of a data condition effective index acquisition method of each production stage provided by the present application;

[0042] Figure 4 is a structure schematic diagram of a feed whole-process production data acquisition system of a smart factory provided by the present application. DETAILED DESCRIPTION

[0043] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined invention purpose, the following describes the specific implementation, structure, features and effects of a feed whole-process production data acquisition method and system of a smart factory according to the present application in combination with the drawings and preferred embodiments.

[0044] Before introducing the specific schemes provided by the embodiments of the present application, the specific processes of the whole production process in the present application are explained to facilitate the understanding of those skilled in the art, and are not limited for the present application.

[0045] The working process of the complete feed production line mainly includes five key processes: cleaning, crushing, mixing, pelleting and packaging, as follows:

[0046] (1) Cleaning process: feed raw materials are cleaned by mechanical equipment to remove impurities, sand and other substances that are not conducive to animal digestion. After cleaning, the raw materials usually need to be dried to remove excess moisture to avoid moisture problems in subsequent processing.

[0047] (2) Crushing process: cleaned feed raw materials are sent to a crusher for pulverization. The purpose of crushing is to process the raw materials into uniform particle size, which helps to improve the efficiency of subsequent mixing and processing. Different types of feed raw materials need to adjust the degree of crushing to ensure the best processing effect and animal digestion and absorption rate.

[0048] (3) Mixing process: various crushed feed raw materials and additives such as vitamins and minerals are mixed according to the preset formula. This process is usually performed by a mixer to ensure uniform mixing of different ingredients and achieve a balanced nutritional composition.

[0049] (4) Pelleting process: after mixing, the raw materials enter the pelleting process, and the mixed feed is pressed into granules. This process not only improves the density of the feed for storage and transportation, but also improves the digestibility and absorption rate of the feed. During the pelleting process, temperature and pressure control are crucial to effectively promote the adhesion of the feed, making it easier for animals to digest.

[0050] (5) Packaging process: the finished feed is automatically packaged through the packaging process. The packaging equipment measures the granular feed, bags it and seals it to ensure that each bag of feed is consistent in weight and the packaging is intact. The packaging material used generally has moisture-proof and rust-proof properties to extend the shelf life of the feed. After packaging, the finished feed is labeled and stored in the warehouse for shipment or distribution to customers.

[0051] In the following description, different "one embodiment" or "another embodiment" refers to different embodiments, which may not be the same. In addition, specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0053] The application provides a feed whole-process production data acquisition method and system for a smart factory.

[0054] Please refer to Figure 1 Fig. 1 shows a step flowchart of a feed whole-process production data acquisition method provided by an embodiment of the application, and the method comprises the following steps.

[0055] In step S100, production data corresponding to each data acquisition device is acquired through different data acquisition devices in each production stage of the feed production process.

[0056] In the feed production process, the precision of the sensor equipment and the layout of the sensor installation position have a significant impact on the effectiveness of data acquisition. In particular, in the multi-formula production scenario, the difference in raw material ratio will cause changes in physical properties such as moisture content and granularity, and further cause problems such as adhesion, accumulation and impact of the transmission pipeline. The representation of these abnormal phenomena on the production data is closely related to the correlation between the sensor equipment positions. Therefore, the data correlation between the sensor equipment positions needs to be analyzed in detail, and the data acquisition strategy needs to be optimized to improve the control ability of the production process.

[0057] Considering that the environmental data and raw material production state data in each production stage have a direct relationship with the control of the quality of the finished product and semi-finished product of the corresponding production stage, different data acquisition devices are set in each production stage of the whole feed production process to monitor the production data of the processing process in each production stage in real time.

[0058] As a specific example, different types of sensors are used as data acquisition devices, the number of sensors is selected according to the length of the production process in different production stages, the production data collected by the sensors in the feed product production process is extracted, and the production data is stored in the order of the production process. The sampling period is set to 1 hour, and the data is collected every 5 minutes. The production data is the data collected by the corresponding sensor, and in order to avoid the influence of the dimension problem on the subsequent data analysis result, each data collected in the embodiment is standardized. The method for data standardization is a known technology, and will not be described in detail here.

[0059] More specifically, in some embodiments, temperature sensors, humidity sensors, flow sensors, etc. can be provided in the production stage of the cleaning process; vibration sensors, current sensors, and pressure sensors, etc. can be provided in the production stage of the crushing process; torque sensors, temperature sensors, and current sensors, etc. can be provided in the production stage of the stirring process; temperature sensors, pressure sensors, humidity sensors, and laser particle size instruments, etc. can be provided in the production stage of the granulation process; weight sensors, visual inspection cameras (for extracting target size data), etc. can be provided in the production stage of the packaging process. The types and locations of sensors are different according to different feed production requirements, and the implementer can determine them according to the specific implementation scenario.

[0060] At this point, the production data of each data acquisition device in each production stage within each preset sampling period can be obtained, and the production status under the corresponding production stage can be monitored in real time. The implementation mainly aims at the correlation and influence between the data collected by different data acquisition devices at different positions in the same production stage, and adaptively adjusts the data acquisition frequency to achieve the purpose of improving the effectiveness of the collected data.

[0061] Step S200, according to the difference between the data monitored by the data acquisition devices at adjacent positions in each production stage, and combining the position information of each data acquisition device in each production stage, the position influence coefficient of each data acquisition device in each production stage is obtained.

[0062] The accuracy of the data acquisition device is easily affected by the different positions of the device. Production problems such as adhesion and accumulation often lead to coordinated changes in data at adjacent positions. For example, a blockage at a certain position will simultaneously affect the data acquisition results at different positions before and after. The difference between the differences between the data collected by the data acquisition devices at two different positions can reflect the real-time situation of the current data acquisition. Is it due to adhesion, accumulation, and other problems that run through the production links of adjacent positions, leading to synchronous data anomalies, or does it reflect the real production process problems, and the difference in data changes due to position.

[0063] Based on this, in any one production stage, in a data collection cycle, first, the discrete degree or fluctuation degree of the production data collected by each position corresponding data collection device is analyzed, and then the difference between the data discrete degree or fluctuation degree analysis results of the adjacent two positions is analyzed, so as to quantify the characteristic analysis result of the first stage, that is, to use the position correlation factor to represent the correlation between the change of the data collected by the corresponding data collection device and the data collection position. Further, combined with the distance between each data collection device and the production position of the production stage, the position influence coefficient corresponding to each data collection device is finally quantified. The closer the data collection position is to the semi-finished product output position of the production stage, the greater the direct influence of the data on the semi-finished product quality is, and the position influence coefficient further comprehensively reflects the influence of the data collection situation of the position where the data collection device is located on the production stage output quality.

[0064] In step S300, according to the similarity between the position influence coefficients of different data collection devices in each production stage, combined with the fluctuation of environmental information in each production stage, the data condition effective index of each production stage is obtained.

[0065] In the whole process of feed production, the pellet quality is highly dependent on the rationality of the production process setting, which is specifically manifested as the synergistic effect of key processes such as cleaning, crushing, mixing, and pelleting. For example, the particle size distribution of corn particles in the crushing stage directly affects the uniformity of the mixing process, and then determines the structural strength and nutrient retention rate of the pelleted product. This multi-stage indirect influence characteristic makes the position setting of data monitoring a key factor in capturing quality fluctuations.

[0066] The direct influence of the process parameters (such as crusher speed, mixing time) of each production stage (such as crushing, mixing) on the semi-finished product quality is dominant, and the installation position of the monitoring equipment (such as the outlet of the crusher and the inlet of the mixer) determines the sensitivity and representativeness of data collection. If the abnormal performance of the monitoring data at different positions in the same production stage is highly similar, it indicates that the abnormality may be caused by the problem of upstream process parameter setting, rather than the influence of the monitoring position itself.

[0067] Based on this feature, first, the similarity between the position influence coefficients of different data collection devices in the same production stage is analyzed to determine the stability degree of the data monitoring position influence in each production stage. Combined with the position influence coefficient of each data collection device in each production stage, the sensitivity of the data monitoring position to the production quality abnormality perception can be more accurately and intuitively reflected, that is, to determine the final abnormal sensitivity factor.

[0068] Further, considering that in the multi-proportion feed production scene of the smart factory, the difference between the equipment operation data and the product quality data is essentially due to the difference in physical properties of the formula raw materials. This difference reflects the effectiveness of data monitoring in the production whole process from multiple aspects.

[0069] For example, the physical properties of different multi-proportion feed formulas interfere with the production process in multiple dimensions. Specifically, increasing high-fat or high-protein ingredients increases the stickiness of the feed, while higher humidity raw materials can cause the feed to absorb moisture after mixing, increasing the risk of sticking. In addition, uneven particle shape and size, as well as compression during transportation, can also increase the contact pressure between particles, leading to sticking. The greater the difference in data obtained during the proportioning process, the more significant the potential impact of proportioning adjustments on the final quality.

[0070] In terms of the contrast relationship between equipment operation data and product quality data, although formula changes will change process parameters such as feeding ratio and mixing time, the setting parameters of core equipment (such as crushers and granulators) are usually constant. This characteristic of fixed parameters and variable inputs allows the equipment operation data to more purely reflect changes in material physical properties rather than fluctuations in the equipment's own state.

[0071] If the abnormal patterns of monitoring data for different formulas in a production stage are highly similar, it indicates that the data monitoring location setting in that production stage cannot effectively distinguish formula-specific abnormalities, and the monitoring value needs to be re-evaluated. For a specific formula, such as high-stickiness feed, the monitoring data for a certain stage may have significant distinguishing degrees, and the monitoring strategy needs to be optimized independently for that formula, such as increasing the sampling frequency.

[0072] Based on this feature, during the analysis of the effectiveness of data monitoring in each production stage, the data monitoring process corresponding to different feed formulas during feed production is further introduced. It can be understood that each feed formula corresponds to a complete feed production process, i.e., the analysis process of the abnormal sensitivity factor in each production stage corresponding to each feed formula is the same.

[0073] Based on the analysis of the influence of each data monitoring location under different feed formulas, combined with the deviation of environmental data in the feed production process, the effectiveness of the data collected by each data monitoring location in each production stage is comprehensively analyzed, and the data condition effectiveness index of each production stage is finally determined.

[0074] Step S400, according to the data condition effectiveness index of each production stage, analyze whether to adjust the data collection frequency, and determine the data collection method of each data collection device.

[0075] For each production stage, the greater the value of the data condition validity index, the greater the degree of effectiveness of the data collected in the corresponding production stage, indicating that real-time monitoring of the condition of the feed product production process is of great significance under the corresponding data collection frequency of the data collection device set in the production stage, reflecting that the data collection parameter setting is more reasonable at this time, so it is not necessary to continue the real-time monitoring process by adjusting the data collection frequency too much.

[0076] The smaller the value of the data condition validity index, the smaller the degree of effectiveness of the data collected in the corresponding production stage, indicating that data monitoring at the current data collection frequency has certain deficiencies, such as underestimating the quality impact or not capturing environmental differences, at which time it is necessary to further increase the data collection frequency to capture the detailed features reflected by the data in more detail, and to improve the data collection frequency to effectively monitor the abnormal condition of the production process.

[0077] Based on this, when the data condition validity index of each production stage does not meet the effectiveness condition, the data collection frequency of the data collection device of the production stage is adjusted and then data collection operation is performed; when the data condition validity index of each production stage meets the effectiveness condition, data collection operation is performed at the current data collection frequency. The effectiveness condition is specifically that the data condition validity index of the production stage is greater than or equal to a preset effective threshold.

[0078] The data condition validity index of the production stage does not meet the effectiveness condition means that the degree of effectiveness of the data collected in the production stage does not meet the requirement, and the data condition validity index of the production stage meets the effectiveness condition means that the degree of effectiveness of the data collected in the production stage meets the requirement. As a specific example, considering that the value of the data condition validity index is a normalized value. In this embodiment, the value of the effective threshold is 0.2, which can be set by the implementer according to the specific implementation scene.

[0079] When the data condition validity index of each production stage is smaller and less than the effective threshold, it indicates that the degree of effectiveness of the data monitoring result is low at this time, and the data collection process or data collection strategy needs to be adjusted, and the embodiment adjusts the data collection frequency to capture the detailed features of the production process in more detail. The specific adjustment operation can be determined by relevant professional personnel. Different levels of data collection frequency can also be set, and when the data condition validity index of the production stage is less than the effective threshold, the data collection frequency of the data collection device in the production stage is increased by one level, and the higher the level, the higher the data collection frequency, and the frequency interval between adjacent levels is the same.

[0080] When the data condition validity index of each production stage is greater and greater than or equal to the validity threshold, it indicates that the data acquisition strategy at this time is more appropriate, and the data acquisition result can effectively monitor the production process, so it is not necessary to adjust the data acquisition frequency at this time, and the production process can be continuously monitored.

[0081] In some embodiments, as shown in FIG. 2, the method for obtaining the position influence coefficient in step S200 can be implemented by steps S201 to S203. Figure 2

[0082] In step S201, a position correlation factor of each data acquisition device in each production stage is obtained according to the difference between the fluctuation of the production data corresponding to each data acquisition device in each production stage and the fluctuation of the production data corresponding to the data acquisition device at the adjacent position.

[0083] In this embodiment, an arbitrary production stage is taken as an example for illustration. First, the monitoring difference degree of each production data corresponding to each data acquisition device is determined based on the difference between each production data corresponding to each data acquisition device and the mean value of all production data.

[0084] It can be understood that in step S100, a plurality of production data in a data acquisition period is obtained by each data acquisition device, and then the deviation between each production data and the mean value of the overall production data can be monitored, reflecting the data difference degree corresponding to each production data. The monitoring difference degree of an arbitrary production data of an arbitrary data acquisition device in a data acquisition period can be specifically represented as wherein X i,n represents the nth production data of the ith data acquisition device in an arbitrary production stage in a data acquisition period, represents the mean value of all production data of the ith data acquisition device in an arbitrary production stage in a data acquisition period.

[0085] Second, the position correlation factor of each data acquisition device in each production stage is determined based on the difference between the mean value of the monitoring difference degree of all production data corresponding to each data acquisition device and the mean value of the monitoring difference degree of all production data corresponding to the data acquisition device at the adjacent position.

[0086] ​It can be understood that the data deviation degree of each production data corresponding to each data acquisition device is calculated, which reflects the data deviation of each production data, and then the mean value is used to represent the data deviation of the overall production data of the data acquisition device in a data acquisition period, and then the difference between the data deviation of the data acquisition device and the data acquisition device at the adjacent position is measured to determine whether the performance of the data deviation between the data acquisition device and the data acquisition device at the adjacent position is consistent.

[0087] More specifically, the embodiment calculates the absolute value of the difference between the mean value of the monitoring deviation degree of all production data corresponding to each data acquisition device and the mean value of the monitoring deviation degree of all production data corresponding to the data acquisition device at the adjacent position, and normalizes the calculated absolute value of the difference to obtain the position correlation factor corresponding to each data acquisition device. The normalization method can be maximum-minimum normalization, and the implementer can also select according to the specific implementation scenario.

[0088] When the difference between the deviation degrees of the production data corresponding to the two adjacent data acquisition devices is larger, it means that the characteristic fluctuations of the data corresponding to the two positions have large difference, and the difference of the data fluctuation is more likely to be caused by the different positions. When the difference between the deviation degrees of the production data corresponding to the two adjacent data acquisition devices is smaller, it means that the characteristic fluctuations of the data corresponding to the two positions are consistent, and the difference of the data fluctuation is more likely to be caused by the influence of adhesion, accumulation, impact and other conditions between different positions.

[0089] Therefore, the position correlation factor represents the correlation between the data fluctuation of the data acquisition device at different positions and the position distribution. The larger the value of the position correlation factor of the data acquisition device, the greater the influence of the abnormal performance between the data collected at the corresponding position and the data collected at other positions on the position of the data acquisition device.

[0090] It should be noted that the embodiment mainly analyzes the performance of the data deviation between each data acquisition device and the data acquisition device at the adjacent next position in the production process. For the last data acquisition device, the data acquisition device at the adjacent next position cannot be obtained, and the data characteristic performance between the data acquisition device and the data acquisition device at the adjacent previous position can be analyzed.

[0091] In step S202, the relative distance factor of each data acquisition device in each production stage is determined based on the distance between the position of each data acquisition device in each production stage and the output position of the corresponding production stage.

[0092] The closer the distance between the location of any data collection device and the location of the semi-finished product output, the more the data performance at the location of the data collection device can be approximated to directly affect the quality of the semi-finished product. The farther the distance between the location of the data collection device and the location of the semi-finished product output, the more the data performance at the location of the data collection device can be approximated to directly affect the quality of the semi-finished product.

[0093] In this embodiment, the number of data collection devices contained between the location of each data collection device in each production stage and the output location of the corresponding production stage is taken as the corresponding relative distance factor. For example, the relative distance factor corresponding to the data collection device closest to the output location of the production stage is 1, that is, the data collection device closest to the output location of the production stage and the output location contain itself.

[0094] In other embodiments, the implementer can also select other ways to determine the relative distance factor according to specific implementation scenarios, such as determining the actual distance between each data collection device and the output location of the production stage in the production process as the corresponding relative distance factor, and the like.

[0095] Step S203, determining the location influence coefficient of each data collection device in each production stage based on the ratio between the location correlation factor and the relative distance factor.

[0096] Specifically, for any production stage, the ratio of the location correlation factor and the relative distance factor of each data collection device is determined as the location influence coefficient of each data collection device. The location influence coefficient comprehensively considers the direct influence of the monitoring location and the spatial distance, and represents the significant degree of the influence of the data monitoring location of the data collection device on the quality. The greater the value, the greater the influence of the data monitoring location of the data collection device.

[0097] In some embodiments, as shown in FIG. 3, the method for obtaining the data condition effective indicator of each production stage in step S300 can be implemented by steps S301 to S304. Figure 3

[0098] Step S301, clustering the data collection devices of the same production stage based on the similarity between the location influence coefficients of different data collection devices in each production stage, and obtaining the category fluctuation coefficient of each production stage according to the aggregation of the data collection devices in the corresponding clustering result of each production stage.

[0099] ​Specifically, for any one production stage, the Euclidean distance between the position influence coefficients of each two data acquisition devices is used as a classification measurement distance to perform density clustering on all data acquisition devices to obtain a clustering result corresponding to each production stage, which contains one or more clusters. The density clustering can use a DBSCAN clustering algorithm, which is a known technology, and the specific implementation method will not be described in detail here. In other embodiments, the implementer can also select a suitable density clustering algorithm according to the specific implementation scenario.

[0100] More specifically, in one production stage, the position influence coefficients corresponding to all data acquisition devices form a data set as sample points, reflecting the influence intensity distribution of different data monitoring positions. The data acquisition devices corresponding to position influence coefficients with similar values are grouped into the same cluster by a density clustering algorithm. The data acquisition devices in the same cluster are located at data monitoring positions that have similar influence on the output quality.

[0101] Further, in order to measure the stability of the data monitoring positions of all data acquisition devices in each production stage on the output quality, the class fluctuation coefficient is obtained by: for each production stage, obtaining the proportion of the number of data acquisition devices contained in each cluster in the clustering result as the class proportion of each cluster in the clustering result, and taking the dispersion degree of all class proportions in the production stage as the class fluctuation coefficient of the corresponding production stage.

[0102] For any one production stage, if the position influence coefficients of most data acquisition devices are close in value, it means that the data monitoring positions of most data acquisition devices have similar influence on the output quality, which is further reflected in the data clustering result that most data acquisition devices are distributed in one cluster, and the number of data acquisition devices contained in other clusters is small.

[0103] In this embodiment, for any one production stage, the ratio between the total number of data acquisition devices contained in each cluster in the clustering result and the total number of data acquisition devices contained in the production stage is taken as the class proportion of each cluster in the clustering result, and the variance of the class proportions of all clusters corresponding to the production stage is calculated to obtain the class fluctuation coefficient of the production stage, that is, the dispersion degree of data is measured by variance.

[0104] When the influence degrees of most data acquisition devices are similar, in the clustering results of the corresponding production stage, the proportion of the category of a certain cluster is large, and the proportions of the categories of other clusters are small, and thus the variance of the category proportions of all clusters is large, and the value of the category fluctuation coefficient is large. When the influence degrees of the data acquisition devices in the production stage are different, in the clustering results of the corresponding production stage, the values of the category proportions of the clusters are close, and thus the variance of the category proportions of all clusters is small, and the value of the category fluctuation coefficient is small.

[0105] Therefore, the category fluctuation coefficient of the production stage reflects the stability and difference of the influence degrees of the positions of the data acquisition devices on the output quality.

[0106] In step S302, based on the position influence coefficient and the category fluctuation coefficient of all data acquisition devices of each production stage, an abnormal sensitivity factor of each production stage is determined.

[0107] In this embodiment, the ratio between the mean value of the position influence coefficient and the category fluctuation coefficient of all data acquisition devices of each production stage is taken as the abnormal sensitivity factor of each production stage.

[0108] The larger the value of the category fluctuation coefficient is, the higher the consistency and the greater the similarity of the influence degrees of the data monitoring positions are, and thus the influence modes of the data monitoring of different positions in the production stage are highly similar, and to a certain extent, the data monitoring positions of the production stage cannot effectively distinguish, and the sensitivity of the data acquisition result to the data anomaly is low. Meanwhile, the smaller the mean value of the position influence coefficient of the production stage is, the lower the sensitivity of the data monitoring positions to the position influence to perceive the data anomaly is, and thus the value of the abnormal sensitivity factor is smaller.

[0109] The larger the value of the category fluctuation coefficient is, the lower the consistency and the smaller the similarity of the influence degrees of the data monitoring positions are, and thus the influence modes of the data monitoring of different positions in the production stage are highly different, and to a certain extent, the sensitivity of the data acquisition result to the data anomaly is high. Meanwhile, the larger the mean value of the position influence coefficient of the production stage is, the higher the sensitivity of the data monitoring positions to the position influence to perceive the data anomaly is, and thus the value of the abnormal sensitivity factor is larger.

[0110] In step S303, environment data of the positions of each data acquisition device in each production stage under different production parameters is obtained to form an environment data sequence of each production stage under different production parameters.

[0111] In the embodiment, different production parameters refer to different feed ratios, and the environmental data can include temperature data and humidity data. The environmental data of the position of each data collection device is sequentially recorded according to the order of the feed production process. In order to avoid the influence of dimension problems on the data analysis results, each environmental data is also subjected to standardization processing. At the same time, considering that data monitoring is performed for each production stage with a complete data collection period, the average value of the environmental data of the position of each data collection device within a data collection period is taken as the value of the environmental data sequence element.

[0112] That is, for any production stage, for each data collection device, the average value of all temperature data within a data collection period and the average value of all humidity data are calculated, respectively, as the environmental data of different types of the data collection device. Then, according to the order of the production process of the production stage, the temperature data and humidity data of all data collection devices are sequentially arranged to form the environmental data sequence of the production stage.

[0113] It should be noted that the data monitoring process under each production parameter, that is, under each feed ratio, is exactly the same, and the environmental data sequence of each production stage in the production process of each feed ratio can be obtained by the same method. For the same reason, the abnormal sensitivity factor of each production stage in the production process of each feed ratio can also be obtained by the same method as steps S301 and S302.

[0114] In step S304, the difference between the environmental data sequence of each production stage under different production parameters and the balance of the environmental data sequence of the same production stage under all production parameters is obtained, combined with the abnormal sensitivity factor, to obtain the data condition effective index of each production stage.

[0115] First, based on the average value of the environmental data of the same data collection device at the position of each production stage under all production parameters, the environmental balance sequence is determined.

[0116] In order to provide data basis for subsequent analysis of the environmental influence of different feed ratios in the production process of the same production stage, it is necessary to first determine the balance of the corresponding environmental data under all feed ratios, which represents the typical environmental state under different feed ratios. For example, for any production stage, an environmental data is obtained at a data collection device under a production parameter. The average value of all environmental data of the same type at the i th data collection device under all production parameters is calculated as the environmental balance parameter of the i th data collection device, and then the environmental balance parameters of all data collection devices in the production stage sequentially form the environmental balance sequence. It can be understood that the types of environmental data include temperature and humidity.

[0117] Secondly, based on the ratio between the difference distance between the environment data sequence of each production stage and the environment equilibrium sequence and the abnormal sensitivity factor of each production stage under all production parameters, the data condition effective index of each production stage is determined.

[0118] In this embodiment, for any one production stage, the DTW distance between the environment data sequence of the production stage and the environment equilibrium sequence is calculated for each production parameter, which measures the deviation between the actual production environment and the overall benchmark state, excludes the interference of environmental fluctuations on the quality influence, and the larger the deviation value, the more significant the environmental fluctuations. As a specific example, the data condition effective index of any one production stage can be specifically represented as:

[0119]

[0120] wherein SC k represents the data condition effective index of the kth production stage, ZY k,m represents the abnormal sensitivity factor of the kth production stage under the mth production parameter, i.e. the feed ratio, S k,m represents the environment data sequence of the kth production stage under the mth production parameter, S k represents the environment equilibrium sequence of the kth production stage, DTW(S k,m , S k ) represents the DTW distance between the sequences, M represents the number of types of feed ratio, i.e. the number of types of production parameters, Norm is a normalization function, and ε0 is a preset minimum value hyperparameter, which is (0, 0.1), used to avoid the extreme value of the DTW distance being 0 to affect the result of data calculation, i.e. to avoid the denominator being 0, which can be taken as 0.01 in this embodiment.

[0121] For each production stage, the data monitoring results of the data acquisition device corresponding thereto can not only sensitively respond to abnormal conditions of the quality, i.e. the larger the value of the abnormal sensitivity factor, but also resist environmental interference, i.e. the smaller the value of the environmental deviation, at this time the effectiveness of data acquisition is higher, i.e. the larger the value of the corresponding data condition effective index.

[0122] The data condition effective index of each production stage can comprehensively reflect the effectiveness of the data monitoring process of all data acquisition devices in the production stage.

[0123] In some embodiments, as shown in Figure 4 , a feed whole-process production data acquisition system of a smart factory is also provided, which is used to implement the steps of the feed whole-process production data acquisition method of the smart factory, and the feed whole-process production data acquisition system of the smart factory specifically comprises:

[0124] a data collection module, configured to acquire production data corresponding to each data collection device through different data collection devices in each production stage in the feed production process;

[0125] a position feature analysis module, configured to obtain a position influence coefficient of each data collection device in each production stage according to a difference between data monitoring of production data corresponding to adjacent data collection devices in each production stage, and combined with position information of each data collection device in each production stage;

[0126] an effective feature analysis module, configured to obtain a data condition effective index of each production stage according to a similarity between position influence coefficients of different data collection devices in each production stage, and combined with fluctuation of environment information in each production stage;

[0127] a collection method determination module, configured to analyze whether to adjust a data collection frequency according to the data condition effective index of each production stage, and determine a data collection method of each data collection device.

[0128] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A feed whole-process production data acquisition method for a smart factory, characterized in that, The method comprises the following steps: Obtaining production data corresponding to each data acquisition device through different data acquisition devices in each production stage of the feed production process; According to the difference between the data monitoring of the production data corresponding to the data acquisition devices at adjacent positions in each production stage, combined with the position information of each data acquisition device in each production stage, the position influence coefficient of each data acquisition device in each production stage is obtained; According to the similarity between the position influence coefficients of different data acquisition devices in each production stage, combined with the fluctuation of the environmental information in each production stage, the data status effective index of each production stage is obtained; According to the data status effective index of each production stage, whether to adjust the data acquisition frequency is analyzed, and the data acquisition method of each data acquisition device is determined; According to the difference between the data monitoring of the production data corresponding to the data acquisition devices at adjacent positions in each production stage, combined with the position information of each data acquisition device in each production stage, the position influence coefficient of each data acquisition device in each production stage is obtained, specifically including: According to the difference between the fluctuation of the production data corresponding to each data acquisition device in each production stage and the fluctuation of the production data corresponding to the data acquisition devices at adjacent positions, the position correlation factor of each data acquisition device in each production stage is obtained; Based on the distance between the position of each data acquisition device in each production stage and the output position of the corresponding production stage, the relative distance factor of each data acquisition device in each production stage is determined; Based on the ratio between the position correlation factor and the relative distance factor, the position influence coefficient of each data acquisition device in each production stage is determined; According to the difference between the data monitoring of the production data corresponding to the data acquisition devices at adjacent positions in each production stage, combined with the position information of each data acquisition device in each production stage, the position influence coefficient of each data acquisition device in each production stage is obtained, specifically including: Based on the similarity between the position influence coefficients of different data acquisition devices in each production stage, the data acquisition devices of the same production stage are clustered, and the category fluctuation coefficient of each production stage is obtained according to the aggregation of the data acquisition devices in the corresponding clustering result of each production stage; Based on the position influence coefficients and category fluctuation coefficients of all data acquisition devices in each production stage, the abnormal sensitivity factor of each production stage is determined; Obtaining environmental data of the position of each data acquisition device in each production stage under different production parameters to form an environmental data sequence of each production stage under different production parameters; According to the difference between the environmental data sequence of each production stage under different production parameters and the environmental data sequence of the same production stage under all production parameters, combined with the abnormal sensitivity factor, the data status effective index of each production stage is obtained.

2. The feed whole-process production data acquisition method of a smart factory according to claim 1, characterized in that, The position correlation factor of each data acquisition device in each production stage is obtained according to the difference between the fluctuation of the production data corresponding to each data acquisition device in each production stage and the fluctuation of the production data corresponding to the data acquisition device at the adjacent position, and specifically includes: For each production stage, the monitoring difference degree of each production data corresponding to each data acquisition device is determined based on the difference between each production data corresponding to each data acquisition device and the mean of all production data. The position correlation factor of each data acquisition device in each production stage is determined based on the difference between the mean of the monitoring difference degrees of all production data corresponding to each data acquisition device and the mean of the monitoring difference degrees of all production data corresponding to the data acquisition device at the adjacent position.

3. The feed whole-process production data acquisition method of a smart factory according to claim 1, characterized in that, The category fluctuation coefficient of each production stage is obtained according to the aggregation of the data acquisition device in the clustering result corresponding to each production stage, and specifically includes: For each production stage, the category proportion of the data acquisition device contained in each cluster in the clustering result is obtained as the category proportion of each cluster in the clustering result, and the dispersion degree of all category proportions in the production stage is taken as the category fluctuation coefficient of the corresponding production stage.

4. The feed whole-process production data acquisition method of a smart factory according to claim 3, characterized in that, The data status effective index of each production stage is obtained according to the difference between the environmental data sequence of each production stage under different production parameters and the environmental data sequence of the same production stage under all production parameters, combined with the abnormal sensitivity factor, and specifically includes: The environmental equilibrium sequence is determined based on the mean of the environmental data at the position of each data acquisition device under all production parameters for each production stage, and the data status effective index of each production stage is determined based on the difference distance between the environmental data sequence of each production stage under all production parameters and the environmental equilibrium sequence and the ratio relationship between the abnormal sensitivity factor.

5. The feed full-process production data acquisition method of a smart factory according to claim 4, characterized in that, The abnormal sensitivity factor of each production stage is determined based on the position influence coefficient and the category fluctuation coefficient of all data acquisition devices in each production stage, and specifically includes: The ratio between the mean of the position influence coefficient of all data acquisition devices in each production stage and the category fluctuation coefficient is taken as the abnormal sensitivity factor of each production stage.

6. The feed whole-process production data acquisition method of a smart factory according to claim 1, characterized in that, The data acquisition method of each data acquisition device is determined by analyzing whether to adjust the data acquisition frequency according to the data status effective index of each production stage, and specifically includes: When the data status effective index of each production stage does not meet the effectiveness condition, the data acquisition frequency of the data acquisition device of the production stage is adjusted and then the data acquisition operation is performed; when the data status effective index of each production stage meets the effectiveness condition, the data acquisition operation is performed at the current data acquisition frequency.

7. The feed whole-process production data acquisition method of a smart factory according to claim 6, characterized in that, The effectiveness condition is specifically that the data status effective index of the production stage is greater than or equal to a preset effective threshold.

8. A feed whole-process production data acquisition system of a smart factory, characterized in that, The system is used to realize the steps of the feed whole-process production data acquisition method of the intelligent factory according to any one of claims 1-7, and the feed whole-process production data acquisition system of the intelligent factory specifically includes: The data acquisition module is configured to acquire production data corresponding to each data acquisition device through different data acquisition devices in each production stage of the feed production process. The position feature analysis module is configured to obtain a position influence coefficient of each data acquisition device in each production stage according to a difference between data monitoring of the production data corresponding to the data acquisition devices of adjacent positions in each production stage and position information of each data acquisition device in each production stage. The effective feature analysis module is configured to obtain a data condition effective index of each production stage according to a similarity between the position influence coefficients of the different data acquisition devices in each production stage and a fluctuation of environment information in each production stage. The acquisition method determination module is configured to analyze whether to adjust a data acquisition frequency according to the data condition effective index of each production stage and determine a data acquisition method of each data acquisition device.

Images